Remember when texting became the thing and we stopped calling each other?
Phone calls carried tone. Hesitation. The sigh that told you more than the words did. Texting flattened that, but at least there was still a human on the other end, responding in real time. Then came Slack. Then async everything. Now we're routing decisions through AI systems that only read what's written down. And we're calling it efficiency.
We didn't just lose the sigh. We lost the instinct.
I was in a design review not long ago. A client looked at the concepts on screen and said they were "still thinking about the style."
They weren't still thinking. They hated it.
The pause. The way they said it. The body language that arrived half a second before the words. "Still thinking" was doing a lot of heavy lifting in that sentence, like when someone says "interesting" about your idea in a meeting and you spend the rest of the day wondering if that was a compliment or a funeral.
None of that was in the meeting notes. None of it would show up in a transcript or a summary. But it was the most important information in the room.
Because of it, we pushed the concept further. Explored beyond what was on the table. We talked about it, then talked some more about it. And the direction that came out of those conversations was the one that worked.
If that meeting had happened over Slack, or been routed through an AI, "still thinking about the style" is all we'd have. Clean. Documented. Off.
This is not an article about AI.
It's about what humans need from each other to make any of it work.
Right now, the entire conversation around AI-native work is about human-to-machine context. How to prompt better. How to build knowledge bases for agents. How to give AI enough background to execute well. We are getting very good at this.
What we are not designing for is humans transferring context to each other. That gap, quiet, unspectacular, slightly awkward and higher stake, is the one that determines whether the work is good and can actually scale beyond one person.
This is about reimagining collaboration. Not the tools. The neurons firing and transferring.
Every Human Carries Three Types of Context
This isn't a leadership problem. It's a human one. Every person in any organization is walking around with three layers of context at all times. Most of us have never named them. Well, let's name them now.
Explicit Context Type 1
Explicit context is what's documented. The ticket. The spec. The roadmap. The recorded meeting. Transferable by systems and especially fit for AI. The layer organizations invest most heavily in capturing, and the one that gives the false impression that context has been shared.
Latent Context Type 2
Latent context is what could be said but never was. The reasoning behind the decision. The budget conversation that happened in the hallway three weeks before the ticket was written. The option that got ruled out and why. It can be articulated, someone just never thought to. It masquerades as explicit because the output looks complete. The spec looks thorough. It's like a beautifully decorated cake with no actual cake inside. Everything checks out until someone takes a bite and realizes there's a hole.
Tacit Context Type 3
Tacit context is what can only be felt through proximity to a full human being. The pause after "still thinking." The sigh after "it's fine." Pattern recognition built from watching the same people in the same rooms across dozens of decisions. But also: the cultural reading of a silence. The instinct built from navigating systems that weren't designed for you. The warmth that makes a client trust you enough to show their real reaction instead of the polished one. This is lived experience as a professional asset. You can't write it down. You can't prompt for it. You can only teach someone to use their spidey senses, as my three year old says, and only if they're in the room with a full human being, not a flattened version of one.
Think of it this way. Explicit is the recipe. Latent is why your grandmother added that extra pinch of salt because it was too plain last time. She knew, she just never wrote it down. Tacit is why her version still tastes different even after she told you everything she did. Well, because she's grandma.
The Gap Nobody Named
Organizations invest in systems for explicit context. Some invest in mentorship programs for tacit context. Almost nobody has a process for latent context, because nobody named it until now.
That's where most handoffs actually fail. Not because the documentation was bad. Because the reasoning that felt obvious in the room never got said out loud to the right person.
The new person reads the ticket. Goes into the workshop. The executive says "it's fine." Takes it at face value. Six weeks later, priorities shift. The decision quietly unravels. Nobody can explain why.
You gave them the map. Not the territory.
Different Context, Different Transfer
Each type needs a completely different approach.
Explicit — document it. Systems handle this. Async is fine. This is the solved problem.
Latent — narrate it. This requires a human deciding to say the thing that felt obvious in the room. "Here's why we didn't do X." "The budget conversation that shaped this spec." Add the "why not" to the ticket alongside the "what." Most onboarding covers what was built. Almost none covers what was ruled out. That's where latent context lives, in the gap between the decision and the reasoning.
Tacit — be in the room. Co-presence over time. The new person needs to watch the client say "still thinking" while you're still there to show them what it means. You can't transfer pattern recognition. You can only teach someone where to look. That requires overlap. Real time. A full human being on both ends of the conversation. And this is where having a more junior person in the room is so valuable, not only because organizations need to train the entire workforce on tacit context, but also because in the act of showing, there's learning.
Start Before the Transition
You don't have to be wrapping up a project or a feature set to start doing this.
Narrate while you work. Add the "why not" to the ticket, not just the "what." Think out loud in working sessions and capture them by transcripts. Make your reasoning visible while you still hold it, because it feels obvious now and invisible the moment you leave the room.
Your real IP is not the output. The output can be reverse-engineered from the Figma file or the Linear board. The reasoning behind it can't. And the read on the room? That lives nowhere but in you, until you find a way to pass it on.
The X-Shaped Bottleneck
Here's where it gets specific.
X-shaped people, those who operate across product, design, technology, and go-to-market simultaneously, tend to hold all three context types at once, across multiple functions. The user insight that shaped the engineering tradeoff. The sales conversation that quietly moved the roadmap. The budget constraint that never made it into the ticket. The read on the executive in the room when they said "it's fine."
That's not just more context. It's higher-density context, layered across domains, held in fewer people.
In an AI-native environment, this matters more than it ever has. AI can now execute at speed. Entire workflows that used to require coordination across teams can be orchestrated by one person with the right tools. That's the promise. But execution only scales if the humans around it can speak to each other.
A specialist leaving an org creates a skill gap. An X-shaped person leaving creates a comprehension gap, across multiple functions, simultaneously. Not because they're slow, but because things move incompletely without them. Most orgs don't realize this until the person is already gone. That's why there's usually a celebration, and a freak out, when an MVP leaves the team.
We are spending enormous energy designing for human-to-machine context transfer. Teaching AI to read the room. Building systems that remember, synthesize, and act.
It's time to design for human-to-human context transfer with the same intention.
The future of collaboration isn't just about what we hand to AI. It's about what we show and tell each other. And that requires something no system has figured out how to replace: a full human being, in the room, paying attention.
#FutureOfWork #FutureOfCollaboration #XShaped #BuildInPublic #DesignLeadership #ProductLeadership #AITools #HumanValue
This is Part 7 of an ongoing series on X-shaped people and the future of innovation teams. Read Part 1: The X-Shaped Individual: Solving for Problems in 3D and Part 2: It's Only Through Doing That You Become: How X-shaped people are made — and how teams can grow them and Part 3: Your Design System Isn't a Style Guide Anymore — It's AI Infrastructure and Part 4: How Do We Price Human Judgment When 5 Hours Turns Into 30 Minutes and Part 5: Architecting Adoption: How Design Systems Survive Like a Language and Part 6: When the Truth Isn't Absolute: Managing Variance Across Design, Code, and Docs
Thu Do is a hands-on product owner with 10+ years bringing products from 0-to-1 across startups, Fortune 500 consultancies (BCG, PwC), and innovation studios. She helps early-stage to early-growth companies ($1-10M ARR) and innovation teams turn big visions into competitive market-ready products and services through human-centered design, product alignment, and AI innovation. This article originally appeared on Thu's Tech Dialect. Find her on LinkedIn.
Co-created with ClaudeCo-created with Claude