AI Won't Replace You. Your Lane Will.
Three phases every company will go through on AI adoption journey — and the one most people are already too late to wait for.
Nobody in your leadership meeting will say this out loud:
Most companies aren’t blocked by AI capability. They’re blocked by what they built before AI existed — and the quiet shame of admitting their foundation was never designed to be readable by anyone but the person who wrote it.
This is Phase One. And if your company is older than a year, it’s probably where you’re stuck too.
Phase One: Before AI Can Help You, It Needs to Understand You
There’s a version of AI adoption that looks great in a board deck and does almost nothing in practice. You buy the tools. You run the pilots. The code suggestions come back wrong, the reviews miss context — and someone in leadership concludes AI just “isn’t ready yet.”
It is ready. Your codebase isn’t.
AI is only as useful as the signal you give it. And most engineering orgs have spent years optimizing for shipping, not for legibility. Variables named tmpFlag. Monoliths where one function touches six domains. No tests, no documentation, no clear contracts between services. That’s the invisible wall between you and every productivity gain you’ve been promised.
Phase One is about fixing this — modularizing, writing tests AI can read, building the eval pipelines that tell you when AI output can be trusted.
And I want to be clear: this work — eval frameworks, platform engineering, database scaling, observability layers — is some of the most intellectually demanding work in the industry right now. It won’t be automated away. It’s what makes automation possible for everyone else.
Phase One isn’t a phase you skip. It’s the phase that decides whether Two and Three are even possible.
Phase Two: The Scoreboard Changes
For the last decade, the unspoken scoreboard in engineering has measured one thing: who builds the most, fastest. Lines of code, tickets closed, features shipped. That scoreboard is changing.
In Phase Two, the people who used to say “I can’t code” start shipping. PMs build internal tools. Designers create working prototypes. Marketing experiments with landing pages without waiting weeks for an engineer.
And engineers? They become the taste layer — the people who know what good looks like, who can see when AI-generated code is technically correct but architecturally wrong.
I’ve seen engineers who were chronically behind — always one sprint short — become star performers after seriously adopting AI. Not because the work got easier. Because AI removed the friction hiding their judgment. They had it all along. They were just spending too much time wrestling with syntax to show it.
The engineers who struggle in Phase Two are the ones who built their identity around output rather than perspective. That anxiety is pointing at something worth examining — not because your skills are obsolete, but because you might be holding on to a version of your value that was always smaller than what you’re capable of.
Phase Two doesn’t shrink engineering. It surfaces what engineering was always really about.
Phase Three: No Lanes. Just Problems.
A team of five. No “that’s not my lane.” No two-week wait for a resource to free up. Just people with a shared problem, approaching it from every angle — building, writing, analyzing, iterating — simultaneously.
Not because everyone’s a unicorn. Because the cost of switching contexts has collapsed. You can write the code, draft the positioning, prototype the UX — not perfectly, but well enough to move, then hand it off for refinement.
Phase Three companies ship in days what used to take quarters. They don’t have role confusion — they have role fluidity.
This isn’t a distant future. It’s already the operating model of the best startups being built right now. The question is whether your company gets there on purpose, or scrambles to catch up later.
Why I Stopped Waiting
I left a career that was working — the title, the comp, the trajectory. Not because I was unhappy, but because I could see that the companies that matter in five years are being built in Phase Three mode today. I didn’t want to spend years helping a large org navigate Phases One and Two when I could try to build Three from scratch.
Joining an early-stage startup was a bet on skipping the line — not out of impatience, but out of conviction.
What I didn’t expect: after years in management, I thought my IC days were behind me. Turns out they weren’t. They were just waiting for the right leverage.
This Is Personal, Not Just Organizational
The hard part of this transformation is never the technology.
It’s identity.
We build our professional self-image around our specialty. I am the engineer. I am the designer. I am the strategist. Those identities gave us clarity, belonging, a sense of what we’re for. Phase Three asks you to loosen your grip — not abandon your depth, but stop treating it as a fence.
This is exactly why I dialed up my coaching practice. The people I want to work with are at the inflection point: engineers who feel the ground shifting, managers wondering what leadership looks like when everyone is becoming a generalist, ICs who sense they have more to offer but haven’t found the operating model that lets them show it.
Navigating this isn’t just about learning new tools. It’s about being willing to rewrite the story you’ve been telling yourself about what you’re worth.
The window to get ahead of this — rather than just catch up — is open right now.
But windows close.
If you’re wondering what the next move looks like for you — I’d love to think through it together.