I was sixteen, maybe seventeen, working the shipping operation for a small publishing house. My job was simple: box the books, run the postage, get the orders out. The books were about process improvement and quality control — dense, technical stuff aimed at engineers and operations managers I’d never met and couldn’t have imagined.
In the downtime between orders, I taught myself to juggle, and I started reading the books almost immediately. Out of curiosity, and a restlessness I didn’t yet have a name for.
But somewhere between the shipping manifests and the packing tape, I absorbed a way of thinking about systems, variation, and process that I had no professional vocabulary for yet — and wouldn’t for years.
I wasn’t actively learning about process control. I was just in the room with books about stuff, right? That distinction turns out to matter more than I understood at the time.
There’s a concept in complexity theory called the adjacent possible. Stuart Kauffman introduced it to describe the boundary of what can come next in biological evolution — not all futures are accessible from a given state, only the ones one step away. Steven Johnson later applied it to innovation: you can’t leap to arbitrary ideas; you can only reach ideas that neighbor your current knowledge and experience.
I didn’t encounter this framework until long after I’d already lived it several times over. When I did, it named something I’d been carrying without language.
That’s the thing about the adjacent possible. From my experience and in the patterns I’ve observed across organizations, you rarely recognize you’re moving into it in real time. You’re just following the problem, the curiosity, the need. The new territory only becomes visible in retrospect.
Most Organizations Try To Leap
Most organizations approach AI adoption the way people approach career planning before they’ve lived enough to know better: they try to leap.
The pattern is familiar enough to be a cliché, but familiarity hasn’t made it less common. Leadership identifies the destination — “we’re getting AI” — and announces it org-wide, often with namedropped companies and models, and almost always paired with vague assurances that this won’t result in layoffs. Then comes the silence. Eventually, links appear, and maybe some basic instructions. Rarely anything that draws a clear line between existing strategy, the gaps AI is meant to address, and how the workforce is actually supposed to use the tools toward that end.
What happens next is predictable. The futurists and early adopters — the people who are curious and adaptive by nature — will engage. They’ll explore, experiment, build local workflows. But their work won’t be aligned or focused, because no one told them what problem they’re solving for the organization. Everyone else pokes at the tool with a stick, gets nothing useful back, and walks away. Adoption flatlines. Leadership concludes the tools weren’t as useful as advertised.
The tools were fine. It’s that steps were skipped.
Don’t Skip Steps
You can’t adopt an AI-native workflow if your team doesn’t share a vocabulary for how AI systems work. You can’t build that vocabulary if people don’t feel safe admitting what they don’t know. You can’t build that safety if leadership is signaling — through rushed timelines, vague reassurances, and the absence of any real curriculum — that uncertainty is a problem to be hidden rather than a condition to be managed.
Each step is only accessible from the one before it. Skipping them doesn’t accelerate the journey. It just means you’ll retrace them later, at higher cost, after something has already broken.
The organizations that move differently do something structurally simple but operationally demanding: they communicate a phased plan that connects strategy to tooling, provides a curriculum that serves every level of expertise and every function, and gives people immediate next steps they can take today. The longer arc includes touchpoints — moments to come together, review what’s been learned, and focus further with more clarity on goals. The plan evolves as the org learns.
The result isn’t that everyone becomes an AI power user overnight. It’s that everyone understands where this is going, how it connects to their work, and what their next step is. Everyone comes along for the ride — or at minimum, knows enough to make an informed decision about how they engage with it.
That’s not a technology problem. That’s an instructional design problem. And it’s solvable.
Closing The Gap Deliberately
Knowing the adjacent possible exists is half the work. The other half is building the bridge to it.
This is where instructional design enters — not as a corporate training concept, but as a discipline for closing the gap between where people are and where they need to be. The best practitioners of it start not with curriculum but with analysis: what does current capability actually look like, what does the next state require, and what specifically stands between the two?
That gap is the unit of work. Not “we need to learn AI” but “our teams can’t evaluate model output because they have no shared criteria for what good looks like.” Instead of the old trope “designers need to code,” the better lens is “designers need enough systems literacy to have a productive conversation with an engineer about tradeoffs.”
In a healthcare UX org, that analysis might produce a capability map showing that clinical researchers can evaluate AI-generated synthesis but lack the vocabulary to specify what they need from a model — which points to the first curriculum artifact being not an AI tools tutorial, but a shared glossary.
The adjacent possible tells you where you’re going. Instructional design tells you how to get there without skipping the steps that make arrival possible.
Organizations that combine both — that map the next achievable state and build deliberate learning infrastructure to reach it — don’t just identify the gap. They close it on purpose, with a plan the whole org can see and follow.
How I Know This Is True
I was trained as a designer — classically, at art school — and somewhere in my first decade of professional work I drifted into engineering. Not because I planned to. Because the problems I was solving kept pulling me one step further into territory I didn’t have a name for yet.
Design gave me the vocabulary for how things should feel and function from a human perspective. Engineering gave me the vocabulary for why things behave the way they do at a system level. Neither one was a leap. Each one was adjacent to where I already stood.
In college, for gas and grocery money, I took a job caring for three disabled adults in their home. Long shifts — sometimes a full weekend, 24 to 48 hours at a stretch. It was unglamorous work in the way that most essential work is unglamorous.
Caring for someone who is wholly dependent on you — for meals, mobility, dignity — compresses your awareness in a specific way. You stop taking your own capability for granted. You develop a constant low-level inventory of what you have, what they need, and what the gap is. Not as a burden, but as a discipline. A practice of noticing.
The adjacent possible of that work wasn’t a job skill. It was an orientation — a habit of accounting for the advantages you’re operating with instead of moving through the world blind to them. Gratitude, yes, but the kind that makes you more capable: functional, not decorative.
Neither transition was planned. Both were adjacent to where I already stood.
What This Has To Do With You
If you’re navigating a career transition, a role change, or an org-wide transformation right now — and in 2026, most people in design and technology are doing at least one of those things — the adjacent possible is a more useful frame than almost anything else I know.
The question isn’t: where do I want to end up? The question is: what is genuinely one step away from where I am right now? What capability, relationship, or context sits just at the edge of my current reach?
That’s the step worth taking. Not because ambition is wrong, but because the steps you skip have a way of showing up later as gaps you can’t explain.
I spent time as a teenager running the shipping operation for a small publishing house — boxing and sending out books about process improvement and quality control before I had any professional context for what those books contained. I spent weekends in college learning what it means to be responsible for someone else’s wellbeing, happiness, and safety. I moved from design into engineering without a plan to do so.
None of it was linear. All of it was adjacent.
Originally published on LinkedIn on May 11, 2026.
Hello. I’m Richard Allan Lee. 
