• When AI writes your communication and AI summarizes it on the other end, no human-to-human exchange actually happens. We are performing a theater of communication.

• Communication carries signals beyond information: effort, presence, care, identity. AI transfers information but breaks these signals. That is why AI-polished messages feel hollow even when they read well.

• Not all communication needs you in it. Separating what is informational (let AI handle it openly) from what is relational (needs your presence) protects trust and relationships.

The Communication Theater We All Perform

You have a client presentation due Monday. It is Friday. You have been pushing it all week, handling the urgent, and now it stares at you from your task list. You almost know what should be inside, but you are not fully there. You don’t feel like sitting with that uncomfortable, itchy feeling of getting your thoughts together, figuring out what to say, wrestling with your inner doubt about whether your thinking is right.

So you take a shortcut. You prompt the presentation into existence. Maybe you edit a slide or two. You send it before end of day, claiming it carries your analysis and recommendations.

Your client receives it Friday evening but leaves it for Monday. Monday morning hits with back-to-back meetings, planning, aligning. So they ask AI to summarize your presentation. They skim the bullet points. Close enough.

You both just performed a communication theater. You claimed you wrote what you didn’t write. They claimed they read what they didn’t read. AI talked to AI. No human-to-human exchange took place.

You might think this is an edge case. It isn’t.

This is everywhere

You scroll LinkedIn and three posts in a row read like they were written by the same person. They weren’t, but they might as well have been. A colleague’s email arrives with that unmistakable polish, too smooth to be them. A client report lands on your desk, fifty pages of impressive nothing. Like a Potemkin village of expertise: looks like a city of knowledge, but nobody on the team actually internalized the data or did the thinking.

Over half of long-form LinkedIn posts now show signs of AI generation. Engagement dropped 40% last year. People are walking away from content that doesn’t feel human. Research backs up what you already sense: people perceived as using AI are evaluated as less cooperative and less trustworthy.

The quality of language went up. Grammar is better. Structure is cleaner. And yet something feels off when you are on the receiving end. It doesn’t have to be obvious AI slop. Even polished, human-edited AI text carries a certain absence. You can feel it, but can’t always name it. Something is missing.

What communication actually carries

Writing to someone was never just about transferring information.

When you write something yourself, you carry signals that go beyond the words. Your effort signals care: I spent time on this, on you. Your imperfection signals authenticity: this is really me. Your attention signals respect: I was present with your message. Your perspective signals identity: this is how I see it, not how a language model averages it. Your willingness to put rough thinking out there signals trust: I am putting myself out there.

Think of receiving a beautiful greeting card. You open it. Inside: nothing. No signature, no messy ink, no personal message. The card is perfect. But it carries zero proof that anyone was actually there.

AI can transfer information. But it cannot authentically carry effort, presence, vulnerability, or identity. You can’t signal effort when there was no effort. You can’t signal presence when you were absent from the process.

Does it matter to you that I spent the last two weeks working on this issue? Why?

We are all doing this to each other

You hate receiving AI-generated communication. You feel dismissed by it, bored by it, sometimes insulted by it. When you sense no human is behind it, something in you disengages. Even synthetic comments, whose entire purpose is to start a human conversation, feel pointless. If the comment is generated, what is the conversation for? To feed an algorithm?

But it doesn’t stop you from prompting a post when you are tired. Or letting AI draft that email because you don’t have time. Or polishing a report with AI because the deadline is here and your thinking isn’t.

Two-thirds of professionals use AI for communication. Less than half trust it. We know this isn’t working, and we keep doing it. The ease and convenience of AI lures us into doing to others exactly what we resent receiving. We are victims and perpetrators of the same theater.

We use AI to appear more competent, more articulate, more knowledgeable than we actually are.

The irony is that the more polished we get, the less of us is left. We disappear.

The signals that make communication human, that make it ours, get stripped away one AI-polished message at a time.

Where this goes

If you don’t like receiving AI-mediated communication, others don’t like receiving yours either. Multiply that across every email, every report, every message in your organization. What happens to trust when the signals we relied on for centuries are getting faked, but not well enough for people not to notice?

And so the arms race begins. AI generates communication that feels off. So we use better AI to make it feel more authentic. We prompt it to include typos, to mimic our speaking patterns, to sound more human. Recipients get better at detecting it. So we use even better AI. We are now competing to fake authenticity. And nobody wins because the finish line keeps moving.

The way through

Not all communication needs you in it. Some communication is informational: data, status updates, summaries, logistics. Let AI handle that, openly. No pretense. No theater.

But some communication is relational. It builds trust, carries your thinking, signals who you are, and that you care. That needs you. Even if your version is shorter, rougher, and less polished. Especially then. Because shorter, rougher, but real beats long, polished, but hollow.

For some people, AI makes communication possible where it wasn’t before. Non-native speakers, people with disabilities. That is different from making communication convenient, where effort was the point.

Remember the Friday presentation? What if instead of prompting the whole thing and pretending it was yours, you sent two things: a one-pager with AI-compiled data and context, clearly labeled as such. And a half-page with your actual thinking, your interpretation, your recommendation. The first page is for their AI to process. The second page is for them.

No theater. No pretense. And your client gets something actually useful: your judgment, not a language model’s average.

What this means for your team

Now imagine this dynamic playing out across your entire organization. Every team, every meeting follow-up, every project update. What kind of culture are you building when nobody is actually talking to each other?

I will explore the organizational side of this in a future issue. For now, the question is personal: what are you going to do about your own communication?

Try This

Before you hit send on your next message, ask yourself one question: Does this need me, or just information exchange?

If it just needs information, let AI handle it. Be transparent about it.

If it needs you, write it yourself. Even if it is imperfect. Especially if it is imperfect.

If you read last issue, you will recognize this as the same calibration instinct applied to communication specifically.

The goal is not to stop using AI for communication. The goal is to stop pretending AI-to-AI is human-to-human.

A Final Thought

I catch myself doing the same thing I described above. Prompting a post because I am tired. Letting AI smooth out an email because I don’t feel like struggling with the words. Then feeling that familiar hollowness when I hit send, knowing the recipient will get my information but not me.

We invest so much energy in crafting perfect images of ourselves, yet when it comes to choosing who to trust and work with, we value authenticity above all. We can sense when someone is real, even if their reality is messy or imperfect.

AI supercharges the pattern that corporate culture has been building for decades: appear more knowledgeable, more competent, more thoughtful than you truly are. But now the gap between who we are and who we pretend to be grows faster than ever.

I don’t have a clean solution. But I know every time I choose to write something myself, even when it is harder, even when it is rougher, I stay in the exchange. I stay present. And that is worth protecting.

– Paweł

P.S. If this resonated, I am exploring how AI-mediated communication affects teams and organizational culture in a future issue.

Subscribe now

Sources & Further Reading:

• Scientific Reports (2023). Artificial intelligence in communication impacts language and social relationships. Study showing people perceived as using AI responses are evaluated as less cooperative, less trustworthy, and more dominant.

• Originality.AI / Fast Company (2024). How LinkedIn opened the door to AI slop. 54% of long-form LinkedIn posts showed signs of AI generation; engagement with AI-generated content dropped 40%.

• KPMG (2025). Trust, attitudes and use of AI: A global study. Global study finding 66% of people use AI but only 46% trust it.

• Your brain experiences borrowed capability as your own. You can’t distinguish between “I did this with AI” and “I can do this”.

• How you engage with AI determines whether it amplifies or erodes your capacity, not whether you use it at all.

• One calibration question guides conscious AI use: How much of ME does this task require: my thinking, my review, or just completion?

Hello friends,

Have you finished a project with AI assistance and felt proud of the output but slightly hollow about your role in creating it? Or used AI to speed through work, then struggled to explain your own thinking when someone asked follow-up questions?

That dissonance between productivity and ownership is a signal worth listening to.

Today, I want to give you something practical to resolve it: a framework for using AI in ways that amplify your thinking rather than replace it. I distilled months of research on how AI can amplify or erode our thinking, and 1.5 years of my daily use, to one core question that will guide your conscious AI use.

But first, you need to understand why you can’t trust how you feel about your AI use.

Your brain can’t tell the difference

Every interaction with AI creates tension: productive struggle (which builds capacity) versus AI assistance (which feels smooth and productive). Most people drift toward assistance without noticing because thinking is hard and AI is effortless.

The danger is invisible while it’s happening.

Your brain experiences borrowed capability as your own. You complete tasks with AI help, which triggers the same satisfaction signals as genuine competence. Your nervous system can’t distinguish between “I did this with AI” and “I can do this”.

Have you used AI to complete a complex task, then found yourself unable to do similar work when AI wasn’t available? That’s not incompetence. That’s your brain’s inability to separate AI’s capability from yours during the original task.

Have you read an AI summary of research and felt like you understood it, then struggled to explain the concepts to a colleague? Your brain mistakes AI’s fluent output for your own comprehension.

Have you felt accomplished after completing more tasks faster with AI, without noticing your capacity to do quality independent work was shrinking? Productivity feels like growth, but volume doesn’t mean your capacity is growing.

This isn’t a personal failing. Our brains prefer cognitive ease over effort. Automation bias (trusting AI too readily) makes us favor what AI proposes. Cognitive offloading (letting AI think for us) makes us delegate thinking when tools are available. AI amplifies this naturally – it’s fluent, confident, and persuasive even when wrong. The combination makes AI use feel smooth and productive, while solo work feels harder by comparison.

The typical cycle: You use AI to speed through work. Something feels off, hollow, not quite yours. You promise to use less AI next time. Time pressure hits. You reach for AI again. The cycle repeats. You wake up when you realize doing the work without AI has become difficult or impossible.

You think you’re maintaining capability. But reasoning, judgment, and originality are quietly eroding.

So how do you escape this cycle? The answer isn’t avoiding AI. It’s changing how you engage with it.

How engagement mode determines outcomes

AI doesn’t inherently amplify or erode your capacity. How you engage with it makes the difference.

The same person using the same AI on the same task can produce opposite outcomes. You can use AI to think better or to avoid thinking. The choice happens in how you structure the interaction.

Amplifying engagement means AI makes you think better. It extends your capability, builds understanding, preserves or develops capacity. Amplifying patterns:

• You think before using AI (don’t go straight to AI)

• You evaluate critically with your expertise and intuition

• You transform AI output significantly (not just minor edits)

• You use AI to challenge and extend your thinking (not replace it)

• You generate foundation or first draft yourself first

These patterns work because YOUR brain does the cognitive work: forming ideas, making judgments, creating connections. AI scaffolds and challenges, but doesn’t replace your thinking.

Eroding engagement bypasses your thinking, prevents learning, and degrades judgment. Eroding patterns:

• You go straight to AI without engaging your thinking first

• You accept AI output verbatim or with minimal changes

• You trust AI without critical evaluation

• You use AI to avoid thinking or struggle

• You let AI generate foundation

These patterns erode capacity because AI does the cognitive work while you merely evaluate output. Your brain is in consumption mode, not creation mode.

Knowing these patterns isn’t enough. When you’re reaching for AI, you need a simple way to choose the right engagement.

The calibration question

Before each task, ask yourself one question:

How much of ME does this require: my thinking, my review, or just completion?

This question does two things. First, it interrupts the automatic reach for AI. Second, it forces you to articulate what actually needs your brain versus what AI can handle.

Your answer determines how you engage with AI, not whether you use it, but how. It maps directly to who does the cognitive work:

• My thinking → Think-First Mode

• My review → Partner Mode

• Just completion → Evaluate Mode

Think-First Mode

You do substantial solo work first: outline, analysis, rough draft. Then use AI to challenge, extend, or polish what YOU created. When AI contributes, you transform it significantly, making it yours.

When to use this:

• High stakes work where quality matters most

• Your unique voice or perspective is the value

• You’re building expertise in this domain

Partner Mode

You seed the direction, AI expands, you evaluate critically. This is a back-and-forth collaboration where you’re rejecting or rewriting 40-60% of what AI suggests. You’re not accepting output, you’re in active dialogue.

When to use this:

• You have enough expertise to evaluate quality

• Balance between speed and maintaining capacity

• The task doesn’t require your unique voice

Evaluate Mode

AI generates a draft, you catch errors and improve the output. You add context AI lacks, verify accuracy, and ensure it meets requirements. You’re confident you can spot problems because you’re already strong in this area.

When to use this:

• Low stakes work where you’re already an expert

• Time pressure requires speed

• Your expertise makes the evaluation reliable

I don’t choose to do hard thinking solely to preserve my capacity. I do it because quality, authenticity, and voice matter.

I bring more of myself when:

• The work must be top-notch, reflecting my best thinking

• My unique perspective or experience is the value

• The output should sound like me, not a generic AI

• I want to own the result as genuinely mine

Watch how this works in practice

Let me show you how I use all three modes within a single task of writing this newsletter.

Step 1: Solo Ideation (Think-First Mode)
I close all AI tools and brain dump thoughts in a note. This struggle builds the cognitive foundation everything else rests on.

Step 2: Research to inform my thinking (Partner Mode)
I use AI to find relevant sources, then I read and synthesize them myself. The reading feeds my brain with ideas.

Step 3: Developing thinking (Partner Mode)
I bring my raw thinking to AI and ask it to shape it with Socratic dialog (question-driven discussion). It forces me to look at the problem from different perspectives, find connections I didn’t know existed, challenge my thinking, and form my thinking more quickly.

Step 4: Structure article (Partner Mode)
I ask AI for article structure and flow options, then I evaluate and choose based on the story I want to convey.

Step 5: Solo Writing (Think-First Mode)
With AI closed, I draft on my own. I sit with the discomfort of not knowing what to say. That’s where my voice emerges.

Step 6: Critical feedback (Partner Mode)
I ask AI to review the draft critically. Then I decide which feedback makes sense and make the edits.

Step 7: Technical polish (Evaluate Mode)
AI acts as an editor for grammar, style, and clarity improvements.

Notice the alternation: Step 1 is Think-First. Steps 2-4 are Partner Mode. Step 5 returns to Think-First. Steps 6-7 shift to Partner and Evaluate. I move through all three modes within one task.

Try This

This week: Before each AI interaction, pause and ask:

Does this need my thinking, my review, or just completion?

Choose your mode based on that answer, then follow that mode’s pattern, especially the alternation rhythm. Observe whether the work felt different when you consciously chose how to engage.

When you’re not sure which mode to choose, default to Think-First. Better to do extra solo work than to delegate thinking you needed to do yourself. You can bring AI in mid-task, but you can’t reclaim cognitive engagement after outsourcing it.

Success looks like this: Your cursor is in ChatGPT’s or Claude’s text box. You’ve typed a prompt asking AI to draft something. You pause. You realize you’re about to delegate thinking you actually need to do yourself. You close the window and open your notebook instead.

That pause is success. That’s you maintaining the capacity that makes you valuable. Experiment, refine your calibration over time, and develop intuition for which mode serves which goal and workflow.

A Final Thought

While working on this material, I experienced how AI amplifies not my thinking but my counterproductive behavior, and makes it harder to notice and break the self-defeating cycle.

I have serious impostor syndrome that makes me procrastinate on publishing work.

I have a number of ways to avoid the dreaded moment of publication. Do just another round of research with a slightly different angle. Ensure my writing aligns perfectly with current research. Endlessly perfect my concepts. Just empty busyness to feel productive while never delivering.

In December, after my LinkedIn posts got half a million views, I became obsessed with getting this framework right.

I went through three complete attempts to model the dynamics of AI use with layers of dependencies and nuance. Each time, I’d bounce ideas off AI for hours, refining, restructuring, exploring edge cases.

AI made it so easy to keep going. The analysis felt productive. I was working hard, thinking deeply, making progress.

Except I wasn’t. The prompting illusion felt identical to being productive and making progress.

I was stuck in analysis paralysis. I was heading toward burnout.

AI eliminated the friction that would have forced me to stop sooner. Without that friction, I could iterate endlessly, never shipping anything.

If you find yourself iterating endlessly, optimizing perpetually, or analyzing when you should be doing, pause.

AI amplifies what it encounters. It amplified my productive thinking when I used it correctly. But it also amplified my procrastination, making endless iteration feel like diligent work instead of avoidance.

The friction that would have forced me to stop sooner (exhaustion, frustration, running out of options) never came. AI kept offering new angles to explore. I had to create my own stopping point: deciding this was done enough and shipping it.

– Paweł

P.S. I’m interested in bringing this approach to organizations adopting AI responsibly. If your team wants to adopt AI without eroding the cognitive capacity that makes them valuable, let’s talk.

Schedule a call

Sources & Further Reading:

• Research study (2024). Practice With Less AI Makes Perfect. Passalacqua et al. found partial automation leads to 23% better skill development than full automation, workers who stayed in the decision loop preserved both skills and motivation.

• Research study (2023). The Metacognitive Demands of Generative AI. Study showing AI creates metacognitive challenges where users struggle to monitor their own thinking when AI does the cognitive work, leading to overconfidence in borrowed capability.

• Research Policy (2024). Robots, Meaning, and Self-Determination. Nikolova et al. studied automation’s impact across 20 European countries, finding it consistently reduces autonomy and competence satisfaction, core psychological needs for meaningful work.

• AI creates a daily tension: productivity now versus capability long-term

• Five patterns drive you toward AI: trusting its answers too much, preferring polished output over rough thinking, reaching for AI before trying yourself, getting stuck in AI’s suggestions, and fearing you’ll fall behind

• One pattern keeps you blind: the false sense that using AI makes YOU more competent, when the AI is doing the thinking

The Productivity-Capability Trap

Last issue I said your professional value lives in creative agency, being in the driver’s seat of your work, not the passenger seat. This issue I promised to expand on what it means to maintain that agency.

But as I researched this, I discovered something. The discomfort many professionals experience with AI isn’t one thing. It shows up differently.

You might experience:

• Work that feels less yours – like you’ve become someone who arranges AI’s ideas rather than creates your own

• Growth that’s stopped – productivity went up but your skills stayed flat

• Uncertain contribution – you can’t confidently defend which parts of your AI-assisted work are yours

• Anxiety about keeping up – pressure to use AI more but you’re not sure how to use it well

• Worry about long-term effects – you’re productive with AI now but concerned about where your thinking will be in two years

One framework that might help make sense of some of this is Self-Determination Theory. It suggests we need three things to feel satisfied in our work: autonomy (ownership, feeling of control), competence (growth, getting better), and relatedness (meaningful contribution, knowing your work matters).

When AI does too much of your cognitive work, it can violate these needs. Work feels like arranging someone else’s ideas (autonomy violation). You’re not growing (competence violation). You’re uncertain if your contribution matters (relatedness violation).

But SDT doesn’t explain everything on that list. Anxiety about keeping up? That’s different, social pressure, competition, fear of missing out. Worry about long-term effects? That’s concern about problems that haven’t happened yet.

What I’m developing is a framework to address both. Here’s what I’m learning.

The Core Tension

Here’s the tension: AI makes you more productive. You can ship more work, faster, with better polish. But that efficiency may come at a cost. When you use AI for thinking you could do yourself, you’re not exercising those mental capabilities. And like muscles, though the brain works differently, capabilities you don’t use may weaken over time.

So you face a choice on every task: Do I think through this myself (slower, harder, but I stay strong)? Or do I use AI (faster, easier, but am I getting weaker)?

Most of us swing between guilt and dependency. We use AI, feel guilty about it, promise to think more independently tomorrow, then use AI again because we need to keep up.

Hidden Patterns That Drive You Toward AI

Understanding these patterns, which overlap and reinforce each other, helps you recognize when they’re making the choice for you.

1. Trusting AI Too Much (similar to Automation Bias)

You trust AI suggestions over your own judgment. You ask AI for analysis, get a confident answer, and accept it without checking it against your own thinking.

Example: You’re writing a strategy memo. AI suggests a framework. It sounds good, so you use it. But you never verified if it fits your client’s actual situation. Research suggests a majority of people accept AI suggestions without independently verifying them.

2. Preferring Polished Output (Fluency Bias)

You judge something as better because it sounds smooth and polished. AI’s output is grammatically perfect, well-structured, confident. Your rough thinking feels messy and uncertain by comparison.

Example: You draft an email. It’s clear but not polished. AI rewrites it beautifully. You think “this is better”. But was the original message actually worse, or did polish fool you into thinking so?

3. Reaching for AI First (Cognitive Offloading)

You reach for external help instead of using your internal thinking. Each time you ask AI before trying to think yourself, you skip the mental exercise.

Example: You face a problem. Your first instinct is “let me ask AI”. But you never gave your brain a chance to work through it. Over time, this reflex may weaken your ability to solve problems independently.I

4. Getting Stuck in AI’s Frame (Anchoring Effect)

AI’s first output is so polished you get locked into its framing. You stop exploring your own ideas because they feel inferior by comparison.

Example: You ask AI to brainstorm solutions to a client problem. It gives you three great options. You evaluate those three, but you never generate your own alternatives. AI’s frame became your frame, and you lost the chance to think beyond it.

5. Fearing You’ll Fall Behind (Loss Aversion)

Your fear of falling behind drives you to overuse AI. “I NEED AI to keep up” becomes the justification for every shortcut. The hype around AI combined with social media’s information bubble effect makes it harder to see the truth.

Example: Your competitor ships faster using AI. You panic. You start using AI for everything, even thinking you used to do yourself. Fear of losing competitive position drives the choice.

The Pattern That Keeps You Blind

Here’s what makes this tension so hard to recognize:

6. The False Sense of Improvement

Using AI makes you FEEL more competent. You’re shipping better work, getting praise, solving harder problems. But the AI is doing the thinking, you might not actually be getting better. Research on automation and skill development suggests this creates what some call “automation complacency”, you feel capable because the system performs well, not because your skills improved.

Example: You use AI to write code for six months. You feel like you’re learning faster, becoming more capable. But when AI goes down for a day, you realize you can’t write the same quality code yourself. You thought you were developing expertise, but you might have been developing dependency.

This is perhaps the most concerning pattern because it prevents you from seeing the other five. When you feel competent, you don’t question whether you’re trusting AI too much, preferring polish over substance, reaching for it reflexively, getting stuck in its frame, or being driven by fear. The false sense of growth can mask the reality of capability erosion.

How These Patterns Work Together

These six patterns, which are interconnected, not independent, create a reinforcing cycle that’s hard to escape:

You trust AI, so you reach for it first. Reaching for it first becomes a habit. Its polished output locks your thinking into its frame. You can’t think beyond what AI suggested. You feel productive, which creates a false sense that you’re improving. That feeling prevents you from noticing your independent capability may be weakening. Fear of falling behind without AI keeps you trapped.

Each pattern strengthens the others. The more you trust AI, the more you reach for it. The more you reach for it, the less you exercise independent thinking. The weaker that muscle, the more you depend on AI. The false sense of improvement blinds you to this cycle. Fear keeps you from breaking it.

The result: You’re more productive but may be less capable. Your work is good but doesn’t feel yours. You’re keeping up but potentially getting weaker, and you might not realize it until it’s too late.

Creative Agency: The Answer

This is where creative agency matters. Remember the chef from last issue facing the food replicator?

His value wasn’t avoiding the replicator. His value was knowing when to use it and when not to:

1. Judgment about when - Is this a special anniversary dinner (cook from scratch) or quick Tuesday lunch (replicator for prep)?

2. The skills to deploy - When cooking matters, he draws on experience, reasoning, trained judgment

3. Taking responsibility - He puts his name on the meal either way

Creative agency isn’t the skills themselves. It’s the position you occupy as the one making decisions, knowing when to deploy your skills versus when to delegate, and taking responsibility for those choices.

But to make those decisions, you need to recognize these six patterns in action. Catch yourself before you trust AI without verifying. Question whether polished output is actually better. Notice when you’re reaching for AI instead of thinking. Resist getting locked into AI’s frame, generate your own alternatives. Recognize when fear is driving the choice. And most importantly, question whether you’re actually getting better or just producing better output with AI’s help.

What Comes Next

I’m developing a framework to navigate this tension. Three phases:

RECOGNIZE (this issue) - Understanding how AI may erode capability through these six patterns.

RESTORE (next issues) - Building core capabilities that make you powerful whether you use AI or not:

1. How you think and decide

• Reasoning from first principles

• Evaluating quality and catching errors

• Knowing when AI output is good vs. plausible-but-wrong

2. Where your ideas come from

• Original perspective grounded in your experience

• Novel connections only you can make

• What makes your work distinctly yours

3. How you improve and stay relevant

• Learning through productive struggle

• Building expertise that compounds

• Growing capacity as AI gets better

These aren’t just defenses against AI overuse. These are your non-automatable professional values. Strengthen these, and you’re prepared for whatever AI becomes. If AI erodes them, you become dependent on a tool you can’t evaluate, and eventually, replaceable.

AMPLIFY (future issues) - Working WITH AI skillfully to multiply your impact while preserving what makes you valuable.

In the next issue, I’ll introduce the basics of working with AI without losing yourself–a system I developed through my own practice.

Try This

Exercise: Catch Yourself in the Patterns

This week, notice when you reach for AI. Before you do, ask:

1. Trust check: Am I about to accept AI’s answer without verifying it against my own thinking?

2. Polish check: Am I choosing AI because the output sounds better, or actually is better?

3. Reflex check: Could I think through this myself first, even just for five minutes?

4. Frame check: If I use AI, will I generate my own alternatives too, or just evaluate what it gives me?

5. Fear check: Am I reaching for AI because I’m afraid of falling behind?

6. Improvement check: Am I using AI so much that I can’t tell if I’m getting better or just getting better output?

Just notice. Don’t judge yourself. Awareness is the first step.

A Final Thought

P.S. This framework: RECOGNIZE → RESTORE → AMPLIFY is foundational to The Augmented Mind book I’m writing, a guide to preserving what makes you valuable as AI gets better.

If you derive meaning and identity from your intellectual work and want to maintain intellectual ownership while using AI strategically, join me in the journey.

Subscribe now

Sources & Further Reading:

• Microsoft Research (2024). The Metacognitive Demands of Generative AI. Study showing AI systems impose constant metacognitive demands—requiring users to monitor and control output quality with every interaction.

• MIT Media Lab (2025). Your Brain on ChatGPT: Cognitive Debt. EEG study revealing AI use reduces brain connectivity during writing tasks, impairs learning, and weakens recall.

• Research study (2024). Performance and Metacognition Disconnect in AI Interaction. Study showing users overestimate their performance by more than AI actually improves it. Nearly 60% of participants blindly trusted AI outputs without verification.

• Harvard Business School (2023). Navigating the Jagged Technological Frontier. BCG study with 758 consultants showing 12-40% productivity gains inside AI’s capability frontier, but 19% worse performance outside it.

• Julia Kołodko (2025). AI and Meaningful Work. Applies self-determination theory to AI automation, examining how removing meaningful tasks violates core psychological needs.

• Your professional value doesn’t live in all your work equally. It lives where you maintain creative agency, where you stay in the driver’s seat.

• This feeling of lost ownership is the real discomfort we face when we over-rely on AI and slip from creator to curator without realizing it.

• The craft isn’t “never use AI” or “always use AI.” The craft is knowing when your judgment matters more and when curating AI’s output is just fine.

Hello friends,

Last week I left you with a question: where does our professional value live when AI can produce instant polished output?

I’ve been trying to answer that question for a while now (here, and there). But finally I found a mental model that tells me WHEN my contribution actually matters, and when I can let AI handle execution without losing anything real.

If you’ve ever felt weird about how much AI helped, I figured out why.

I tasked an AI to search for the most relevant sources on what differentiates human and AI contributions to knowledge work.

But my prompt wasn’t specific enough, and AI went a few steps too far, doing the research synthesis by itself.

The synthesis was good and had some ideas I could build on.

And I felt... uncomfortable.

If I used it, it would be like reporting someone else’s thinking, not forming my own. I felt robbed and devalued. It’s been nagging me for days.

If you’ve ever felt weird about how much AI helped, you know this feeling.

That discomfort isn’t about using AI. It’s about something deeper.

It’s about the difference between creating and curating.

Between being in the driver seat versus being in the passenger seat.

Between doing creative work and arranging someone else’s output.

I felt uncomfortable being a curator of AI’s ideas.

Here’s what I’ve come to understand.

My value doesn’t live in ALL my work equally.

My value lives in the parts where I maintain creative agency.

Being in the driver seat. Deciding. Giving direction. Actually creating.

It’s not about HOW I do it (with or without AI assistance). It’s about WHO is driving.

The truth is: not all your work requires creative agency. Not all of mine does either.

Some work is fine as curation.

Last week I introduced an analogy: the chef in a restaurant of the future, facing the food replicator (like in Star Trek) that produces perfect-looking dishes in seconds.

The chef makes food through a messy, longer process. The replicator creates almost flawless dishes instantly.

How does the chef justify his value?

I think he would say something like:

My value isn’t that I’m better than the replicator at every single task. My value is that I know WHEN my craft matters for THIS specific dining moment.

When you sit down tonight, I assess the context:

• Are you celebrating something important?

• Do you have dietary restrictions the database hasn’t seen?

• Are you a regular who trusts my judgment, or skeptical?

• Do you want perfect execution of a classic, or something surprising?

You’re not paying for molecules. You’re paying for my contextual judgment about when to deploy which faculties of my craft, and my accountability for getting that judgment right.

The replicator can match my PRODUCT when the context is standard. But it can’t match my PROCESS of reading your needs, assessing the situation, and deploying the right response.

When primary creative agency matters (special occasion, novel constraints, relationship-building, differentiation):

• I’m in the driver seat from ingredient selection through plating

• I draw on embodied experience of thousands of meals

• I reason through novel combinations no database contains

• I take full accountability for the experience

• I create something uniquely suited to this moment

When secondary curation is fine (busy lunch service, standard preparation, proven techniques):

• I let the replicator handle the reliable execution

• I focus my creative agency on the touches that matter

• I verify quality with my trained judgment

• I still take accountability for the result

The chef’s value isn’t in avoiding the replicator. It’s in knowing when his judgment matters more than perfect execution.

This is true for your professional value too.

Remember that discomfort I felt when AI did the synthesis? That was the warning signal. Creative agency is the cognitive work that makes you irreplaceable.

The danger isn’t using AI. The danger is losing track of when you’re still driving versus when you’ve moved to the passenger seat.

Knowing when creative agency matters is the first step. The next is knowing where it comes from.

Next week, I’ll share the three sources of irreplaceable professional value you must develop to stay in the driver’s seat.

Try This

Prompt: Where do you actually maintain your creative agency?

Your AI assistant already has a memory of how you work. You can ask it to show you the pattern you might be missing.

Important: Use this in the AI chat or project where you’ve done the most work. A fresh conversation has no data to analyze. If you haven’t used AI regularly for work tasks, this prompt won’t give meaningful results.

Copy this into ChatGPT, Claude, or Gemini (works best if you’ve used it regularly):

You know how I work and what I’ve been working on. Analyze our entire conversation history.

I want to understand where I maintain creative agency versus where I operate in curation mode.

Creative agency means: I’m in the driver seat. I’m deciding, giving direction, actually creating. The thinking is mine, even if I use tools to help execute.

Curation means: I’m arranging, editing, or executing based on someone else’s output (including yours). I’m in the passenger seat. The core thinking came from elsewhere.

Compare these two versions of me based on our conversations:
- What % of my work falls into each category?
- Which category builds my irreplaceable professional value?
- What am I protecting well? What am I at risk of losing?
- One pattern I should change this week

Be direct. Give me the uncomfortable truth, not reassurance.
If I’m losing ground, tell me. If I’m protecting the wrong things, tell me.

Treat the output as a starting point, not truth. The AI only sees what you’ve shown it. It may miss context, invent patterns, or tell you what sounds good. Challenge its conclusions.

I’d love to hear what you discover. Share your results in the comments or send me a DM. What patterns surprised you?

Leave a comment

Mine said editing and research synthesis is where I should be more in the driver seat. Ouch.

A Final Thought

I’m writing about preserving creative agency while using AI to write about preserving creative agency.

The irony isn’t lost on me.

But here’s what matters. I stayed in the driver’s seat. The insights are mine. The struggle was mine.

The ownership is mine.

P.S. Next week I’ll share the three sources where your creative agency comes from, and how to develop them. This is the framework I’m building for the book, The Augmented Mind. If you want to stay in the driver seat of your own work, you’ll want to read it.

For professionals who earn by thinking and want to use AI without losing their thinking quality and intellectual ownership, join me as I figure this out.

Subscribe now

Sources & Further Reading:

• Dell’Acqua, F. et al. (2023). Navigating the Jagged Technological Frontier. Harvard Business School Working Paper 24-013. The BCG study with 758 consultants showing those using AI for tasks outside its capability frontier performed 19 percentage points worse than those working without AI.

• Microsoft Research (2025). The Impact of Generative AI on Critical Thinking. Research on how AI confidence correlates with reduced critical thinking effort.

• Harvard Gazette (2025). Is AI dulling our minds?. On cognitive offloading and maintaining independent thinking capability.

Key Points:

• We say we want transparency about AI use, but we punish those who disclose it, reducing trust by 16-20%. The gap between what we say and what we do reveals something deeper.

• The real problem isn’t disclosure ethics. It’s that we don’t have a clear mental model for where our professional value lives when AI can produce instant polished output.

• Until you can articulate what makes your work different from AI’s, admitting AI use feels like admitting you’re replaceable.

The Transparency Paradox: Why Admitting AI Use Feels Dangerous

I feel uncomfortable disclosing I used AI for my work (by AI I mean generative AI, LLMs).

For some reason it feels like cheating.

I don’t want my colleagues, clients, or readers to perceive me as less capable, or my work as less valuable.

But disclosure creates uncertainty around my real contribution versus AI’s.

Makes me question if I could be replaced by a machine and nobody would notice a difference.

Should I hide that I used AI and avoid people asking uncomfortable questions about my contribution?

Using AI in knowledge work is still a gray area.

The problem isn’t transparency about AI usage.

The problem is we don’t know anymore how to value our work.

AI outputs results so fast and seemingly so perfect that it creates a painful comparison with our own iterative human process.

The AI’s polished language makes the necessary human struggle feel inadequate.

It sometimes feels like “I’ll never reach that level of proficiency”.

This undermines self-esteem and belief in capabilities.

I use AI every day and I get intimidated by it, I notice myself losing confidence in my own abilities.

And I’m in my mid-career with over 20 years of professional experience.

Even when writing a simple email, I upload a draft version to AI to proofread it, or at least that’s what I say to myself.

But in reality, I’m seeking confidence that it’s good enough to send.

Do you also have that?

What happens when we can’t trust our own judgment anymore and need an algorithm for that?

We can’t articulate what makes our work different from AI’s work.

If we don’t know what is the value of our work, we don’t want others to question that, so we hide AI use.

Researchers at the University of Arizona’s Eller College of Management ran 13 experiments with over 5,000 participants across education, business, and creative work.

They found that even though over 90% of people say they want transparency about AI use, they penalize those who disclose it, reducing trust in them by 16-20%.

Trust drops even further if somebody else exposes you after using an AI detector.

The penalties are not equal.

Research on over 1,000 engineers found female professionals face twice the competence drop for AI disclosure compared to male colleagues, 13% versus 6% for identical work.

The gap between what we say and what we do shows up everywhere.

We want transparency but punish those who are transparent.

We want authenticity but design sophisticated prompts for AI to sound like us.

We want transparency but fear consequences.

We’re asking ourselves: What are the rules here? Am I doing this wrong? Will I get caught?

We fear devaluation, that if I admit I used AI, clients or colleagues will think I couldn’t do it alone.

Admitting it feels like I can be replaced because others can prompt too.

On the other hand, we feel conflicted about hiding AI use, which creates cognitive dissonance, guilt, and impostor feelings.

My expertise and experience in innovation and technology seem to be losing relevance as knowledge is being democratized by AI.

AI makes people feel super smart, creates echo chambers, gives illusion of expertise they don’t have.

AI creates false confidence trap.

In a way hiding use of AI is a rational decision to protect your professional reputation when it is not clear how to value our contribution.

Imagine a chef in a restaurant of the future facing the same dilemma with food replicators, like in Star Trek, that produce perfect-looking dishes in seconds.

The chef makes food through a messy, longer process. Replicator output looks more perfect and instant. The client could buy a replicator themselves.

Question: Why should they pay the chef?

What makes the iterative, slow and messy cooking process valuable?

If the chef uses the replicator for stock or prep, should they tell the client?

Will the client think, “Why am I paying chef prices for replicated components?”

The uncomfortable truth: The chef isn’t entirely sure what makes their cooking more valuable than the replicator’s output.

Until they can articulate it, admitting they used the replicator feels dangerous.

This is why transparency feels so threatening to us.

It’s not just about disclosure.

It’s about not having clarity on where our value as knowledge workers lives.

Being bombarded by media messages that we’ll be replaced doesn’t help.

When you can’t articulate what makes your work different from AI’s, admitting AI use feels like admitting you’re replaceable.

You’re not broken. This is genuinely difficult. And you’re not alone.

The transparency paradox isn’t really about disclosure.

It’s about the missing mental model for where professional value lives in the AI age.

Until we have that clarity, transparency will continue to feel dangerous.

That’s the work we need to do first. (And that’s what we’ll explore in the next issue.)

Try This

Exercise: Where does your uncertainty live?

When you feel conflicted about AI use, identify which question troubles you most:

1. The organizational question: Am I clear on policies and norms around AI disclosure in my workplace?

2. The contribution question: Can I articulate what I contributed beyond what AI generated?

Your answer reveals what you need clarity on. Not whether to disclose, but where your uncertainty about professional value actually lives.

A Final Thought

I hide AI use too.

Not always. Not with everyone.

But enough that I feel the same cognitive dissonance I’m describing here.

I’ve spent 20 years in technology and innovation.

I should be comfortable with AI. I use it daily.

And still, I find myself either downplaying how much I relied on it or mentioning it too casually, when the truth is more complex.

I’m writing this newsletter using the AI collaboration workflow I’m developing for my book.

I’m testing these concepts on myself first.

Some sections I wrote entirely alone.

Others I co-drafted with AI.

The research synthesis involved heavy AI use.

I’m telling you this because I’m not preaching from a position of having figured it out.

I’m researching where human value lives in AI-augmented work because I need to understand it for myself, not just explain it to others.

If I can’t figure out how to preserve my thinking quality and authentic ownership while using AI, I have no business teaching others how to do it.

This is my discovery process.

I’m sharing it with you.

– Paweł

P.S. I’m writing a book, The Augmented Mind (working title), about maintaining thinking quality and intellectual ownership when working with AI. This newsletter is where I test the concepts, like this week’s transparency paradox, before they become chapters. You’re getting the ideas first 🤓

For consultants, researchers, and professionals who want to use AI without losing their thinking quality and intellectual ownership – join me as I figure this out.

Subscribe now

Sources & Further Reading:

• Schilke, O., & Reimann, M. (2025). “The transparency dilemma: How AI disclosure erodes trust.” University of Arizona’s Eller College of Management.

• Trusting News (2024). “AI Trust Kit: Transparency Research and Newsroom Best Practices.” September 2024.

• McCown, J. (2025). “The case against AI disclosure statements in education.” Inside Higher Ed, October 2025.

• Noti-Victor, J. (2025). “Regulating hidden AI authorship.” Virginia Law Review, 111(139).

• Acar, O. A. (2025). “Research: The hidden penalty of using AI at work.” Harvard Business Review, August 2025.

TL;DR: Survival requires systematic transformation: strategically retire commoditizing services while building new capabilities, invest in people's domain expertise and judgment, make everyone consultative, and use AI to your advantage.

The time for incremental adjustments has passed. What software agencies need now is profound business model transformation – nothing less will suffice.

Through my collaborative work with Calin Muresan, combining our experiences across software agencies and client relationships, we've identified that agencies need systematic change across six strategic areas. They can't fight against this shift. They must embrace it strategically, repositioning themselves toward areas where human capabilities create distinctive value.

Based on our collaborative workshops, thriving agencies are implementing six strategic moves:

1. Take Stock and Transform Your Portfolio

First, get brutally honest about your current service offerings. Map everything you do across those three value territories we discussed earlier - Diminishing, Stable, and Emerging. Look at your margins, what clients are saying, and how you stack up against competitors.

I'm not suggesting you immediately abandon everything in the diminishing value zone. That would be financial suicide for most agencies. Instead, develop a thoughtful transition plan.

Create a "service retirement schedule" where you actively plan the phase-out of commoditizing services while simultaneously developing new offerings higher in the value chain. It becomes a visual roadmap everyone in the company can understand.

One agency I worked with mapped their services and discovered that 60% of their revenue came from work that would be heavily automated within two years. This honest assessment became the foundation for their successful transformation strategy.

2. Invest in People's Capabilities

This is absolutely critical. You need to develop your team in two parallel tracks:

First, deepen your domain expertise. This happens through immersing people in client industries, targeted training, and sometimes strategic hiring. I've seen success when agencies assign people to learn specific industry verticals rather than just technical specialties. The demand is shifting toward business acumen and domain knowledge.

Second, cultivate true mastery through mentorship. The difference between technical craft (which AI is getting good at) and professional insight (which comes from experience) is significant. Create intentional mentoring structures where your most experienced people can transfer their wisdom. All those insights that never make it into documentation or training materials.

AI disrupts knowledge transfer systems that have operated for generations. Academic research suggests that the skills required in occupations most susceptible to AI influence are changing 25% faster than in other occupations. This highlights the urgency for structured mentorship programs that preserve and transfer professional wisdom.

3. Make Everyone Client-Facing a Consultant

Don't just create a separate 'consulting' department. I made this mistake myself. It gives the appearance of adopting a new service model without genuinely integrating the necessary mindset throughout the entire organization.

What's needed is to spread a consultative mindset throughout your entire organization. Every client-facing person needs to develop business understanding and advisory capabilities. From developers and designers to QA and project managers. This means fostering a cultural shift where they are empowered to:

• Challenge requirements, asking 'why' before 'how' and probing for true business needs, rather than just implementing specifications

• Anticipate future needs and spot strategic opportunities clients haven't yet imagined

• Read the room in client meetings, discerning unspoken concerns and building trust through genuine empathy

• Translate complex technical decisions into measurable business outcomes, explaining not just what was built, but why it matters for the client's bottom line or strategic goals

• Understand the unwritten rules and nuances of specific industries, like healthcare or finance, making them indispensable partners

This cultural transformation takes 12-18 months and requires uncomfortable conversations about changing performance metrics and team structures. But agencies that commit to this process see measurable results. Higher margins, stronger client relationships, and natural differentiation from competitors.

4. Use AI for Your Own Advantage

There's a certain irony here. Even as AI disrupts our industry, it also provides tools to help us adapt. Use AI to enhance your own operations while maintaining the human wisdom that ensures quality.

This creates a powerful dual advantage: you become more efficient internally while simultaneously gaining firsthand experience with AI that builds credibility with clients facing similar transitions.

Don't just layer AI onto existing processes though. That's a common mistake I see. Research from MIT Sloan Management Review emphasizes that this approach yields suboptimal results. Take the opportunity to fundamentally rethink how work gets done. How might AI augment decision-making? How could it improve quality assurance? How might it transform your delivery processes?

I've developed practical tools for this systematic work redesign through my Human-First AI Framework. The Intelligence Audit helps agencies map workflows between Craft (what AI can handle) and Wisdom (what requires human expertise). Combined with the Transformation Potential Map, agencies can prioritize exactly where to integrate AI for maximum impact while preserving human value.

These tools are available as open source - I encourage you to test them, adapt them, and make them work for your specific context. You can access the complete framework and assessment toolkit at LINK.

5. Build Responsible AI Governance as a Competitive Advantage

Given that Accenture reports only 2% of companies have fully operationalized responsible AI practices, there is a massive competitive opportunity in the market. Responsible AI governance means having clear frameworks for data privacy, bias detection, security protocols, and ethical decision-making in AI systems.

When agencies develop these capabilities internally - ensuring their own AI use is transparent, secure, and ethical - they gain credibility to guide clients through similar challenges. Clients facing AI implementation are looking for partners who have already solved these problems themselves.

Developing strong internal capabilities in Responsible AI governance becomes a powerful differentiator that builds client trust and positions you as an indispensable strategic advisor.

6. Consider Fundamental Business Model Shifts

Some agencies are exploring even more dramatic transformations beyond service enhancement. This means shifting from being implementation-focused agencies to strategy-led consultancies that solve business problems first, then recommend technology solutions. Instead of "we'll build what you specify," it becomes "we'll help you determine what actually needs to be built and why."

Other approaches include incubation models where agencies co-create and co-own digital products with clients, or creating internal startups that develop proprietary solutions. These models focus on strategic insights and problem-solving rather than just technical execution, potentially creating higher margins and more distinctive market positioning.

The Leadership Imperative

Here's what nobody tells you about transformation - it's messy, uncomfortable, and full of false starts. The agencies succeeding aren't those with perfect strategies. They're those with leaders brave enough to admit the old model is dying.

Be transparent with your team. Share the challenge and the vision. Celebrate small wins. Most importantly, model the continuous learning you're asking from others.

For those of us in the CEE region, our strong technical foundation gives us a good starting point. But we need to intentionally build new capabilities in business acumen, strategic thinking, and domain expertise.

The cost game that built CEE's tech success is ending. But the next game - built on strategic value, human insight, and AI augmentation - could be even better.

Whatever approach you take, Calin and I have both noticed that being honest about AI's challenges builds more credibility than making grand promises. Clients appreciate straight talk about what's actually possible.

The question isn't whether to transform. It's whether you'll lead the change or be dragged along by it.

The choice, as always, is yours.

— Paweł Jarmołkowicz & Calin Muresan

This series emerged from collaborative working sessions between Calin Muresan and myself, exploring how AI advancement will reshape software service value chains. Calin's insights on agency business models and client relationships were instrumental in developing these strategic frameworks. For CEE agencies ready to move beyond the cost game, the future isn't about competing with AI – it's about creating value that only humans can deliver.

Previous:

TL;DR: While AI masters technical craft, human wisdom from experience remains irreplaceable. Agencies must cultivate critical thinking, creativity, and judgment through mentorship and deliberate practice to stay above average in an AI-saturated market.

I met Mike Dewis, an accomplished opera soloist, who said something that stopped me in my tracks: "Opera remains one of the purest performances left. There's no technology standing between the singer and audience. You can only rely on your craft."

As we talked, Mike explained how he relies on both his craft from thousands of hours of practice and his artistry from lived experience. "The craft gets me to competent," he said. "The artistry makes it memorable."

That conversation crystallized something about the nature of human expertise in the AI era. The distinction between what AI can learn from training data and what humans develop through experience suddenly made sense. This distinction between craft and wisdom holds the key to our future in an AI-augmented world.

The Craft and Wisdom Distinction

Mike's insight revealed a powerful framework that applies directly to software development. AI rapidly masters craft - syntax, patterns, best practices, the technical skills. But wisdom? That stubborn human quality that comes from years of successes, failures, and reflection? That remains ours.

Why AI Settles for Average

Here's what's fascinating about AI systems: they're trained to minimize prediction error across millions of examples, which mathematically favors the most common solutions. Exceptional work is rare in training data, so AI produces competent but conventional results.

Think about it: AI learns from millions of code examples where most are, by definition, average. When the system optimizes to be "most likely correct" across this vast dataset, it naturally gravitates toward functional, reliable, but ultimately conventional solutions. It raises the floor of quality dramatically, but struggles to reach the ceiling.

This isn't a limitation we can easily engineer away - it's built into how current AI systems learn. When you optimize for statistical reliability across massive datasets, you sacrifice the outlier brilliance that comes from human insight and experience.

The Irreplaceable Human Capabilities

Through our discussions, Calin and I kept returning to one question:

The answers form a blueprint for our industry's future:

Critical Thinking - Not just solving problems, but questioning whether you're solving the right problem. Developers can save significant time by questioning requirements before implementing, asking "why" before "how."

Creative Connections - Linking ideas from different domains, finding elegant solutions in unexpected places. The developer who applies gaming psychology to enterprise software engagement, the architect who brings manufacturing lean principles to code design and deployment.

Ethical Judgment - Navigating the gray areas where there's no clear answer. Balancing user privacy with functionality, short-term gains with long-term consequences. These decisions require understanding context that extends far beyond code.

Emotional Intelligence - Reading the room in a client meeting, sensing team tensions before they explode, building trust through difficult projects. No algorithm captures the moment when a client's body language says they're unhappy despite their words.

Strategic Vision - Seeing beyond immediate requirements to anticipate future needs. The architect who designs for scale before anyone asks, the consultant who spots opportunities clients haven't imagined yet.

Cultivating Human Wisdom

How do you develop these capabilities? Not through traditional training. Wisdom grows through:

Intentional Mentorship - Pairing experienced professionals with emerging talent, focusing not on technical skills but on judgment development. Creating formal programs where senior staff regularly share decision-making frameworks and hard-won lessons.

Reflective Practice - Creating space to process experiences. Regular retrospectives focused on decisions, not just deliverables. What did we learn? What patterns emerged? Why did our initial assumptions prove wrong?

Diverse Exposure - Wisdom comes from breadth. Rotate people across industries, problem types, client sizes. Each context adds layers to their judgment.

Responsibility Gradient - Give people ownership of outcomes early, with appropriate support. When you're accountable for business results, not just clean code, you develop different muscles.

The Strategic Imperative

The pattern is clear: as AI handles more craft work, the premium for human wisdom increases exponentially. Agencies that intentionally cultivate this wisdom - through structured mentorship, reflective practices, and diverse experiences - create sustainable differentiation.

Your people are not just resources to be optimized. They're the source of everything AI cannot replicate.

Invest accordingly.

— Paweł Jarmołkowicz & Calin Muresan

In the final fourth article of this series, we'll explore the practical transformation roadmap - specific steps agencies can take to build these capabilities while navigating the transition from cost-based competition to wisdom-based differentiation.

Next: Charting the Course: Strategic Transformation for Software Agencies in the AI Economy
Previous:

TL;DR: As AI automates routine technical work, value is migrating toward strategic insights, complex integration, and human judgment. We map which services are losing value, which remain stable, and where new opportunities emerge.

"Where does value move?" Calin asked during one of our sessions. It's the question that changes everything.

When basic technical implementation becomes commoditized - when AI can generate functional code in seconds, automate testing, and create documentation - the entire economics of software services shifts. But here's what I've learned: this isn't random destruction. It's predictable migration.

Through our workshops, Calin and I developed a framework that makes sense of this chaos. Think of your services as living in three territories:

• Diminishing Value - where AI automation erodes margins and commoditizes skills

• Stable Value - where human judgment and relationships maintain their worth

• Emerging Value - where new opportunities command premium rates

Understanding which territory each of your services occupies isn't academic exercise. It's your evolution roadmap.

What's Losing Value Fast

Three areas are losing value rapidly:

Routine Implementation Tasks are the first affected. Basic CRUD operations, standard UI components, boilerplate code generation - these are AI's sweet spot. When GitHub Copilot can boost developer productivity by up to 75% in specific use cases, when AI generates test cases and automates deployment scripts, the premium for basic coding evaporates.

Cost-Based Advantage follows close behind. Just being cheaper isn't going to cut it anymore. Calin had conversations with agencies in Poland who are finding that their traditional price advantage is quickly evaporating as AI tools level the playing field.

Time-Based Billing Models simply break. How do you justify billing 40 hours for what AI accomplishes in a fraction of that? You don't. Truth is, this model was already flawed - it never really encouraged efficiency. Why finish faster when you're paid by the hour? AI just exposed what was always broken. The entire economic equation needs rewriting.

Consider this: deployment time for enterprise machine learning applications has dropped from months to weeks thanks to AI automation. When Infosys developers generated over 7 million lines of code using GitHub Copilot, the implications became clear - basic code generation is becoming a commodity at unprecedented scale.

What Holds Its Ground

Not everything loses value. These areas remain strong because they require something AI hasn't mastered:

Operational Excellence - reliably delivering quality results - never goes out of style. Even with all the AI tools in the world, someone still needs to make sure everything works together properly.

Complex Technical Problem-Solving remains firmly human. System architecture decisions, performance optimization, debugging complex integration issues - these need pattern recognition born from experience. AI assists but can't replace the intuition that says "something's wrong here" when all tests pass.

Domain Knowledge and Business Context prove remarkably durable. Understanding healthcare compliance, financial regulations, or manufacturing processes - and more importantly, knowing how technology serves these domains - requires human judgment. Clients pay for partners who grasp their industry's unwritten rules.

Trust and Relationship Building might sound soft, but it's rock-solid value. I've heard this from multiple clients: "We can replace coders, but we can't replace partners who understand our business." That trust, built through years of collaboration, surviving failed projects and celebrating successes, isn't downloadable.

Strategic Guidance becomes more valuable as options multiply. With AI tools everywhere, clients drown in possibilities. They need advisors who can say "yes, you could build that, but should you?" This becomes even more critical when AI makes it trivially easy to generate code - the temptation to over-engineer solutions multiplies. This requires business acumen that goes beyond technical capability.

Where New Value Emerges

Here's where I get excited. These emerging areas don't just maintain value - they command premium rates:

AI Strategy and Implementation tops the list. Not just using AI tools, but understanding which problems AI actually solves versus which it complicates. I know agencies charging consulting rates for AI readiness assessments and roadmapping.

Human-AI Workflow Design creates entirely new service categories. Organizations need help designing systems where humans and AI complement rather than compete. This isn't about tools - it's about reimagining how work happens.

Complex Systems Orchestration grows critical as software becomes more distributed and AI-powered. MLOps (Machine Learning Operations) and LLMOps (Large Language Model Operations) expertise commands premium rates. Someone needs to make all these pieces work together reliably.

Responsible AI Governance emerges from necessity. Ensuring AI systems remain fair, transparent, and compliant isn't optional anymore. With only 2% of companies having fully operationalized responsible AI practices, agencies with genuine expertise here differentiate themselves immediately.

Outcome-Based Partnerships flip the entire model. Instead of "we'll build this feature," it's "we'll improve your conversion rate by 20%." Taking responsibility for business results, not just technical delivery, changes everything.

Look at the real results: banks reducing fraud losses by 50% using AI, telecom companies saving $7 million annually through AI-driven knowledge centers. These aren't just efficiency gains - they're transformative business outcomes.

The Pattern Behind the Shift

Step back and the pattern becomes clear: value migrates from technical execution to strategic judgment, from isolated tasks to integrated outcomes, from following specifications to challenging assumptions.

The key insight that emerged from our analysis? AI handles craft - the technical skills anyone can learn. But wisdom - knowing when to break rules, understanding unstated needs, seeing connections others miss - remains stubbornly human.

Consider this: while AI helps product managers save 10-30% of time on routine tasks, the real value emerges when they use that saved time for strategic thinking. The skills required in AI-influenced occupations are changing 25% faster than in other fields. This isn't just about working faster - it's about working fundamentally differently.

The big picture? AI is definitely automating routine tasks, but it's simultaneously creating demand for higher-level human capabilities - strategic thinking, complex problem-solving, domain expertise, ethical judgment, and the ability to effectively orchestrate AI systems.

Looking at your own service offerings, which of these emerging areas aligns best with your existing strengths? What could you realistically start developing in the next six months?

— Paweł Jarmołkowicz & Calin Muresan

In the coming articles of this series, we'll explore:

• Article 3: The uniquely human capabilities that become your competitive edge

• Article 4: A practical roadmap for transforming your agency to thrive in the AI era

Next: The Human Edge: Cultivating the 'Human Wisdom' AI Can't Automate

Previous:

TL;DR: The cost advantage that built CEE's software industry is vanishing as AI commoditizes basic coding tasks. Agencies must urgently identify new sources of value beyond being cheaper than Western competitors.

For nearly two decades, Central and Eastern European software agencies thrived on a simple formula: excellent technical skills at competitive prices. This was our winning hand, the foundation that transformed cities like Warsaw, Bucharest, Prague, and Kraków into tech powerhouses. But today, we're watching that advantage dissolve faster than anyone thought it would.

Calin Muresan and I spent hours dissecting this shift during our recent workshops. During our sessions, Calin introduced a game metaphor that perfectly captures what's happening: software agencies can play different games - knowledge, capacity, operational excellence, or cost. In CEE, we mastered the cost game. We built our reputation on delivering quality work at rates that made Western clients take notice.

That game is ending.

The signs have been visible for years - wage inflation, remote work opening up everywhere, increased competition. But AI has turned gradual erosion into an avalanche. When GitHub Copilot helps developers complete complex tasks 12% faster, when AI generates functional code from natural language descriptions in seconds, when basic testing and documentation become automated - what happens to agencies whose primary value proposition was affordable human coding capacity?

Calin and I see it in every client conversation. Where they once asked "How much per developer?" they now ask "What outcomes can you deliver?" The shift is profound and accelerating.

Empirical data and industry reports are already demonstrating this impact: McKinsey research shows AI delivering 10-30% productivity gains across development tasks - meaning work that once required a full development team might soon need just two-thirds of the resources. Gartner predicts that by 2028, 75% of enterprise software engineers will use AI coding assistants, up from less than 10% in early 2023. These aren't distant possibilities - they're happening now.

A friend who runs an agency in Poland shared their wake-up call with me. A major client started using AI tools internally and cut their development team needs by 60% in just three months. The client didn't need cheaper developers anymore. They needed fewer developers. Period.

But here's the twist: that same agency saved the relationship by pivoting to strategic consulting. They stopped selling developer hours and started selling strategic insight. Helping clients decide what to build, not just how to build it. Today, they charge 40% more per hour even though they bill fewer hours overall. This isn't just one agency's story - it's a preview of our collective future.

Here's the hard truth every agency faces from Kraków to Bucharest: when AI can generate boilerplate code, automate routine testing, create documentation, and even assist with requirements gathering, being 20-30% cheaper than Western competitors isn't enough. When basic coding becomes a commodity, cost advantage evaporates.

This isn't about AI replacing all developers tomorrow. It's about the fundamental economics of our industry shifting beneath our feet. The tasks that junior developers spent years mastering - the bread and butter of many CEE agencies - are precisely what AI automates first. These routine implementations, standard CRUD operations, basic UI components, and straightforward testing scenarios represent significant portions of typical projects.

The implications hit hard across every aspect of our business model:

• Time-based billing breaks when AI accomplishes in minutes what took days

• Traditional pyramidal team structures - many juniors, fewer seniors - lose their logic when AI handles junior-level work

• Geographic arbitrage vanishes when location becomes irrelevant to basic coding tasks

Yet I don't see this as catastrophe. I see it as a catalyst for the maturing of our industry in the region. The same disruption that threatens old models creates space for new ones. This shift restructures the entire value equation. When AI handles routine coding, agencies compete on insight, not implementation. They must find new ways to create and capture value that go beyond commoditized technical execution.

Agencies that recognize this shift early, that move quickly toward higher-value services, will find opportunities that didn't exist before.

The question Calin and I kept returning to: if not cost advantage, then what? Where does value migrate when the old game ends?

We built thriving tech ecosystems on the foundation of cost-competitive excellence. Now we must evolve beyond that foundation. The next game won't be about being cheaper. It'll be about being irreplaceable.

This means agencies that help clients navigate AI strategy, that understand complex business domains deeply enough to guide technology decisions, that can orchestrate human-AI teams effectively. The Polish agency's pivot to strategic consulting isn't an anomaly - it's a preview of where sustainable value lies in our industry's next chapter.

— Paweł Jarmołkowicz & Calin Muresan

In the coming articles of this series, we'll explore what this transformation means:

• Article 2: Where value is migrating in the AI-augmented software industry

• Article 3: The uniquely human capabilities that become your competitive edge

• Article 4: A practical roadmap for transforming your agency to thrive in the AI era

Next: Beyond Code & Costs: Where Value is Shifting in the AI-Augmented Software Industry.

This fear plagued my writing journey here. I've been caught in continuous loops of obsessive perfection, delaying publication, rewriting endlessly to prove my worthiness to have a voice. Every time I write, my stomach tightens and my critical voice echoes: "you're not enough," "this won't be good enough." My perfectionism operates on fuzzy logic: if I do this perfectly, I'll eliminate grounds for judgment, or avoid it entirely by never finishing. It's a self-defeating strategy to avoid pain.

It feels like a straitjacket on my mind, keeping me from expressing my true self. Looking for a way out, I started using LLMs as my personal coach alongside my morning journaling practice. Through this collaboration, I discovered something profound: my perfectionism isn't a personality trait, it's a protection mechanism I unconsciously developed growing up to avoid pain.

The Breakthrough That Changed Everything

This realization did more than help my writing. It showed me technology's transformative potential when approached thoughtfully. Not just as a tool, but as a catalyst for authentic self-expression. AI didn't just help me work faster; it helped me understand myself better and become more authentic.

But here's what troubled me: everywhere I look, I see technology doing the opposite. Social media algorithms exploit our insecurities. AI tools make us feel replaceable. Systems prioritize engagement over our wellbeing. We've reversed the fundamental relationship. We adapt to technology instead of technology adapting to us.

My personal breakthrough sparked a bigger question: what if we intentionally designed all technology to unleash human potential rather than diminish it? This led me to develop principles for what I call Human-First Technology manifesto.

At its core, Human-First Technology means:

• Expand human choices while respecting their autonomy.

• Support human wellbeing rather than exploit our vulnerabilities.

• Strengthen human systems and relationships rather than fragment them.

• Create sustainable value that aligns commercial success with human flourishing.

These aren't abstract ideals, they're born from experiencing technology as both liberator and oppressor.

From Principles to Practice: My Focus on Human-First AI

AI's rapid evolution makes these principles more critical than ever. My work now focuses on Human-First AI, not just as a professional stance, but as a deeply personal mission.

This builds on two decades of ML experience, from neural networks and computer vision research in 2006 through co-creating diagnostic and therapeutic to establishing ML departments at software agencies. My early work in healthcare AI highlighted crucial human elements: fear of replacement, building trust, understanding limitations. These lessons shape my approach today.

I focus on three areas:

• Unleashing Human Potential: Designing AI that catalyzes creativity and critical thinking, moving beyond automation

• Symbiotic Ecosystems: Creating effective human-AI collaboration that amplifies our unique capabilities

• Guiding Leaders: Helping organizations align AI's power with human values for sustainable growth

The Courage to Share

Writing this manifesto, and this post, represents overcoming that initial fear of judgment. The same courage required to confront my perfectionism is what we need to question technology's role in our lives.

My future work will focus on building these healthier relationships with technology, with AI as the dominant theme. Because when we design technology that truly serves humanity, we all have the opportunity to flourish.

It is not perfect but I want to share this with you.

I welcome your thoughts.

For over a decade, I've been searching for my path to professional mastery—trying startups, consulting, and management without finding that deeper connection. I labeled myself a generalist, trying to soothe my remorse from what felt like an inability to commit to mastering one field. It has been frustrating trying to be the jack of all trades but master of none.

Then I met Mike Dewis, an accomplished opera soloist who said something that stopped me in my tracks:

When he speaks, how he articulates and uses his voice makes you want to listen. His words brought in me a feeling of awe for what he represents and longing for true professional mastery. There before me stood Mike, embodied mastery, although I know he himself would deny it, being too modest. Someone who dedicated his professional career to mastering operatic performance that escapes modern technological advancement.

In a world that increasingly values efficiency, speed, and technological shortcuts, there's something powerful about rediscovering the deeper satisfaction and personal fulfillment that comes from mastery—the patient development of skill and understanding over time. This isn't about rejecting technology but rather asking how it might enhance rather than replace our journey toward excellence.

Coming from the tech world, I felt I have something important to learn from Mike. In a technocratic culture fostering impatience, efficiency over mastery, over-reliance on technology potentially leading to a decline in human experience:

What remains essentially human in mastery even as technology transforms our work?

Craft and Wisdom: The Two Dimensions of Mastery

Mike's perspective revealed a powerful framework for understanding mastery—one that transcends opera and offers insight for all of us navigating a technology-driven world. True mastery, I realized, consists of two dimensions:

1. Craft - technical skills that can be systematically learned, and

2. Wisdom - the personal dimension that develops through reflective practice and the integration of knowledge into intuition – tacit knowledge, or experiential wisdom.

To illustrate the difference let's take a look at a couple of examples:

1. The Chef's Palate
   A chef can follow a recipe perfectly (craft) but the ability to taste something and instinctively know what it needs—a touch more acid, a hint of heat, a subtle herb—comes from years of cooking, tasting, and developing their palate (wisdom). No recipe can capture this intuitive sensing that allows them to elevate dishes beyond technical execution.

2. The Diagnostician's Intuition
   A doctor can follow diagnostic protocols and interpret test results accurately (craft), but the ability to notice subtle inconsistencies in a patient's story, recognize patterns that don't fit textbook cases, and sense which questions to ask next comes from years of clinical experience (wisdom). No medical algorithm can fully capture this intuitive sensing that allows exceptional physicians to diagnose complex cases that defy standard categorization.

3. The Developer's Elegance
   A programmer can write code that executes correctly and efficiently (craft), but what distinguishes exceptional developers is their ability to create solutions with an intuitive simplicity and adaptability that anticipates future needs (wisdom). Two developers with identical technical knowledge will produce remarkably different systems based on their accumulated experience with how code evolves and how humans interact with technology.

Technology interacts differently with these two dimensions. For craft skills, technology can provide legitimate shortcuts and accelerate learning. However, the wisdom dimension has an inherently experiential nature that can't be bypassed. The result is a form of 'hollow craft' that lacks the deeper understanding that comes only through experience and reflection.

This Craft and Wisdom framework helps us identify when technology truly enhances our growth versus when it offers false shortcuts around essential experiences. I'll deepen this in the following sections.

This distinction between craft and wisdom isn't entirely new - it resonates with several established frameworks in the study of expertise and mastery:

• Michael Polanyi's "Tacit Knowledge" Concept
  Polanyi's distinction between explicit knowledge (which can be codified and transferred) and tacit knowledge (which can only be acquired through experience) parallels the craft/wisdom distinction. His famous phrase "we know more than we can tell" captures the essence of the wisdom element.

• Donald Schön's "Reflection-in-Action"
  In "The Reflective Practitioner," Schön distinguishes between "technical rationality" (applying learned theories and techniques) and "reflection-in-action" (the wisdom professionals develop that allows them to think while doing). This closely aligns with the craft/wisdom framework.

The Shortcut Illusion: What Technology Can and Cannot Accelerate

Technology offers unprecedented efficiency in developing craft skills - the technical foundation of any domain. This efficiency creates genuine value, through the removal of unnecessary barriers, democratizing access to knowledge, and accelerating learning curves. The risk isn't in these shortcuts themselves, but in our cultural blindness to what they can't accelerate: the development of wisdom.

Technology can genuinely accelerate craft acquisition, but it can't substitute for the wisdom that comes only through experience.

A novice chef can quickly learn techniques through video tutorials, and a beginning programmer can leverage frameworks that handle routine coding tasks. But for wisdom development, technology often attempts to bypass rather than support the essential experiences. The chef app that promises perfect dishes without understanding flavor principles, or the AI coding assistant that generates solutions without building the intuition for elegant design or understanding code – these offer the appearance of mastery without the developmental journey that creates true expertise. The result is a form of 'hollow craft' that lacks a deeper understanding.

Wisdom emerges from integrated experience - not just time served, but reflective engagement that transforms knowledge into embodied understanding. This integration can't be downloaded or automated; it must be lived.

Our cultural shift toward immediate results and constant optimization creates an environment where the patience required for mastery development feels increasingly foreign.

Thoughtful Integration: Technology as Partner, Not Replacement

The most powerful relationship with technology isn't rejection or uncritical embrace, but thoughtful integration - knowing when efficiency serves our development and when it might undermine the deeper engagement necessary for wisdom development. This good judgment itself might be the meta-skill of our age.

Perhaps what we need isn't less technology but more intentional technology - tools designed not just for efficiency but for enhancing the integration that leads to true mastery. What would technology look like if designed explicitly to support the development of human mastery in its fullest sense?

From my experience, here are some approaches to thoughtful technology integration that I've found valuable:

Writing with GenAI

• Craft enhancement: Using AI to improve grammar, suggest word alternatives, and format documents

• Wisdom preservation: Developing initial ideas, thesis statements, and unique perspectives yourself before seeking AI assistance

• Thoughtful integration: Writing first drafts manually to develop your voice, then using AI for editing suggestions that you thoughtfully evaluate rather than automatically accept

Research and Learning

• Craft enhancement: Using AI to summarize articles and explain complex concepts

• Wisdom preservation: Forming your own connections between ideas and developing critical perspective

• Thoughtful integration: Asking AI for explanations, then attempting to teach the concept back to verify understanding

Client Presentations

• Craft enhancement: Using AI to format slides, suggest transitions, and polish language

• Wisdom preservation: Developing your unique approach to storytelling and client relationship dynamics

• Thoughtful integration: Outlining your key insights manually, using AI for production aspects, then customizing to reflect your consulting style

The Irreplaceable Guide: Why Mentors Matter More in the AI Age

In our conversation with Mike, we touched on the crucial role of mentors in the pursuit of excellence, particularly in the arts. Mike emphasized the importance of a mentor's ability to provide constructive criticism and guide the artist towards excellence, even if it means challenging the artist's own preferences.

Unfortunately in the tech sector I come from, having mentors is not a common practice. But I would argue that in the world accelerated by AI advancements, mentorship should become standard in professional development. AI accelerates craft skill but can't guide wisdom development. That's where true mentorship can bridge the gap between skill and wisdom by:

1. Providing personalized guidance on which technological shortcuts are valuable vs. which might undermine development.

2. Offering compressed wisdom through stories and case examples that accelerate pattern recognition when similar situations arise, effectively compressing years of experience.

3. Experience teaches mentors which 20% of effort yields 80% of growth. By directing attention to these high-leverage aspects of practice, mentors create learning shortcuts.

4. The seamless integration of craft and wisdom is difficult to understand conceptually but can be observed in a mentor's work. This modeling provides an embodied example of what integration looks like in practice, giving apprentices a target to aim for.

I have true mentors in my life. Not formal arrangements, but informal relationships with people I highly value and look up to. It is a small group of people I make sure to stay in touch with. On numerous occasions they guided me where I didn't develop my own intuition yet or where just their experience outgrows mine. (Thank you 💚 Aga, Łukasz, Jasiek, Jan, Jinder)

Finding Purpose: From Generalist to Focused Explorer

My conversation with Mike began as research for an article but revealed something deeply personal. As he spoke about opera's pure craft and lifelong dedication, I recognized what had been missing in my own journey - not just skill or knowledge, but purpose. A north star worthy of true mastery.

In the process of distilling this Craft & Wisdom framework, I found unexpected clarity about my own calling: to explore how we can develop healthier, more intentional relationships with technology – creating approaches that enhance human potential while building sustainable value for our collective future.

This resonates with my core experiences - from years developing healthcare solutions to studying exponential technologies at Singularity University. I've always been drawn to technology's potential for good, yet concerned about its unintended consequences. What I lacked wasn't capability but clarity of purpose. This is the purpose worthy of dedicated mastery I've been searching for - one that builds on my past while addressing our collective future.

At this critical inflection point where technology is reshaping human experience, this exploration matters on multiple levels:

• For individuals: Finding balance that expands capabilities without undermining wellbeing and authentic connection

• For organizations: Creating systems where technology augments human strengths with sustainable economic models

• For society: Ensuring technology serves human flourishing through thoughtful governance and values

Sometimes the most valuable discoveries come from unexpected conversations. What began with Mike's reflection on opera ended with clarity about my own path to mastery.

Beyond the Framework: Limitations and Promise

Like any framework, this one also has its limitations:

• The neat division between craft and wisdom exists more on a spectrum in reality, with many aspects of "wisdom" teachable through different methods.

• What we consider intuitive (wisdom) in Western contexts may be explicitly codified in other traditions.

• As AI advances, the boundaries between what technology can and cannot accelerate will likely shift.

• The framework focuses primarily on individual mastery when today's excellence often emerges from collective knowledge and networked expertise.

• Mastery itself evolves alongside technology - transforming what we value and how we develop it.

Despite these limitations, I find the Craft & Wisdom lens valuable for approaching technology with better judgment. It offers a practical way to assess when technology truly enhances our development versus when it might offer appealing but ultimately hollow shortcuts.

My exploration of healthier technology relationships is just beginning, but I'm convinced that our most valuable innovations won't come from blindly accelerating everything, but from carefully choosing what to optimize and what to experience fully. Perhaps true wisdom in the AI age lies in knowing which shortcuts to take and which journeys require the full, unabbreviated path.

I invite you to stay with me in this exploration as we seek technology relationships that enhance our humanity rather than diminish it. Feel free to reach out with any questions or to share your own perspectives on mastery in our technological age.

As a newly assigned delivery manager for that account—just a week into the role—I was still trying to figure out what was what. It was a high-visibility project with lots of pressure, and I needed to quickly understand the situation. I joined the virtual meeting with my new co-manager, who was quite a character, as we were still in the process of aligning our approaches.

As I clicked into that video call, I was genuinely nervous. The project was outside my core expertise, and I'd just discovered there were deadline issues. I needed to wear my manager hat well here, not my expert one.

The meeting started well but deteriorated as we progressed, with my co-manager descending into micromanaging. I could see Orlagh's expression changing on my screen as the collaborative energy visibly drained away. I witnessed the exact moment trust broke between us, but I couldn't fully comprehend why.

When Trust Is Broken

Our professional relationship remained strained after that meeting. Orlagh eventually left the project, and I later left the company, with the unresolved tension lingering in my mind as a professional failure.

A year later, while exploring visual thinking as an innovation tool, I thought of Orlagh, an expert in this area. Despite our history, her recent comment on my post encouraged me to reach out. I had an intuition: what if visual communication could help resolve our past disconnect? When I proposed this approach to address our tension, she embraced the idea immediately.

We planned a 90-minute Miro session to test whether visual thinking could help us reconnect, challenged not just by the methodology but by our emotional history.

Visual Communication as Bridge-Builder

Guided by Orlagh, we began by recapping our intentions: bring healing to the past disconnect and test how visual communication could help us do that. The atmosphere was cautious but hopeful as we prepared to revisit our challenging experience.

We decided to start with Orlagh's designed exercises and finish with mine as the closing one. For our warm-up, we selected animal avatars representing our project roles—Orlagh chose a goldfish while I selected a whale.

I explained choosing the whale—large, a bit clumsy with its own inertia but also protective, using its big body to shield the team from disruptions. As I spoke, something shifted; Orlagh seemed surprised, saying she hadn't realized how much I cared about the team. This metaphor expressed intentions that direct conversation couldn't.

Next, we created a timeline with weather images showing our feelings throughout the project, both marking the disconnect moment with blue stars. Remarkably, we placed our stars at exactly the same moment—no explanation needed. The playful metaphors revealed layers we wouldn't normally discuss, helping me see her perspective while my explanation of context created visible understanding in her.

Our third exercise used basic shapes to represent team structure and relationships, allowing me to explain power dynamics and differing interests in our operating contexts. This discussion of team personalities took us off track. Having used our planned 90 minutes, we agreed to extend by 30 minutes for a final exercise addressing the specific moment of disconnection.

Orlagh's exercises sparked extensive discussions between us about the context we both operated in, constraints and operating models that were imposed, and how these affected us and our work. All these conversations brought better understanding of each other and prepared us for the final moment of truth.

The Moment of Truth

For our final exercise, we each filled in a template with three concentric circles:

• The inner circle was "My Experience"—our thoughts and feelings in that moment

• The middle circle was "My Observations"—what we saw and heard from each other

• The outer circle was "Context"—all those project pressures and external factors

This became the culminating moment where we laser-focused on that one fallout meeting. I struggled with my drawing skills—wishing we also had pre-existing visual elements to express myself more clearly like with Orlagh's exercises. This became the most powerful part of our session. It was uncomfortable, honest, and in the end, truly transformational.

The pivotal moment came when Orlagh explained how in that meeting she realized our stated intention didn't match reality. "I felt betrayed in that meeting," she said simply. Initially, she directed blame toward my co-manager. For a second, I felt a wave of relief—here was an escape path where I wouldn't have to take full responsibility.

But that wouldn't be fair to Orlagh or true to myself.

I lowered my voice and said, "I'm responsible for that. I allowed it to happen. I betrayed your trust." I found myself repeating the words as I processed the admission in my head. Something shifted in Orlagh's expression as I verbally acknowledged my role—a subtle opening, perhaps the first true reconnection since that disastrous meeting.

I recalled planning with my co-manager two agendas for that meeting: the official one everyone knew about, and a hidden one where we wanted to assess Orlagh's leadership and how her team could handle deadline problems.

That was the moment I dropped all my explanations and justifications for what was, clearly, an ill-advised decision. I knew I'd made a choice that ran against how I normally work with others—how I try to lead with transparency and honesty. I should have stayed true to my values. No pressing issue or desire to make a good impression justified compromising my integrity like that.

Looking at her face on the screen, I realized our experiment had worked—visual communication had created a path to reconnection that conventional approaches hadn't managed in a year of distance.

What We Discovered About Visual Communication

We both agreed visual communication enabled different kinds of sharing that don't naturally emerge in dialogue. The visual elements created distance, making it easier to discuss our experience as if it were a story rather than our own difficult reality.

We also noticed how each exercise revealed a different perspective:

• The animal avatars showed our INSIDE view (how we saw ourselves)

• The team shapes exercise revealed our OUTSIDE perspective (how we operated in the system)

• The concentric circles exposed the BETWEEN US dimension (our direct interaction and its breakdown)

The exercises built a foundation of contextual understanding that enabled us to address our underlying tension. The concentric circles exercise finally allowed us to "take the bull by the horns." Separating feelings from observations created clarity that had been missing in our previous interactions.

When Visual Communication Works for Resolution (And When It Doesn't)

While this approach worked for us, important considerations for applying similar methods include:

Requires foundation of trust: This method worked because we knew each other well enough to be vulnerable. You can't simply instrumentalize this approach with just anyone—there needs to be some relationship foundation to build upon.

Timing matters: We approached this retrospectively after a year, with cooled emotions. Fresh tensions might require a neutral facilitator.

Freedom from constraints: No longer working together allowed us to speak freely without workplace politics or ongoing project concerns.

Skills and tools matter: Pre-existing visual elements were crucial, as limited drawing skills would have distracted from expressing thoughts and feelings clearly.

Despite these limitations, visual communication creates a unique space for addressing disconnects that verbal approaches can't resolve, externalizing the tension to enable both objective discussion and deeper emotional connection.

While the most powerful moment came through words—my acknowledgment of betraying trust—the visual journey created the conditions for that honesty. For innovation practitioners, this approach won't replace direct conversation but can create pathways that make those conversations finally productive.

— Paweł