Conversational maps

A conversational map is a taxonomy of a field: the concepts a discipline uses when it talks about its work, collected and laid out so that everyone can see the same landscape. It is not the truth about the field. It is a shared interface, the thing that lets two people with different mental models point at the same part of the work and mean the same thing.

This page is the oldest layer of our thinking. Ola wrote the first version years before dark context had a name, in the middle of a decade spent making card decks that give teams a shared language for their work. It is published here because it explains what the method actually does, and why so much of every field's map turns out to be missing.

Before maps existed, people navigated on mental models. Everyone in the village knew how the house related to the well and how the meadow related to the hilltop. They shared a language for the landscape, often specific to that one valley. Nothing was drawn, because for the people inside the landscape nothing needed to be.

Every organization is such a village. How the work actually gets done lives as a shared mental model: who to ask, what matters, why the rule exists. It navigates fine, and it is written down nowhere. Nobody inside notices, because insiders are exactly the people who never need the map. There is a name for knowledge held this way: tribal knowledge.

Mobility. When people began leaving the known landscape, orientation had to travel with them. The map was invented as shared understanding for strangers: a frame that lets people with different heritage and different mental models align, like two people glancing at wristwatches that show the same time.

Organizations meet their strangers all the time. A new hire, a partner, a consultant: each one gets months of guided wandering, which is what onboarding is. And now there is a new kind of stranger, an AI that has read every book and visited nothing. A general model knows every landscape in general and yours not at all. The stranger has arrived, and the village still has no map.

Listen to how people talk about their work, and write down what keeps coming back. Ask practitioners what they do, what matters, and why. Keep track of the order in which things appear. Some elements are mentioned once and never again; others repeat through every conversation, and the repetition is the signal. Those are the core concepts of the field.

Try it on a restaurant. Staffing, menu, interior, lighting, food type, branding, opening hours, payment options, target group, marketing, guest experience, signage, reviewers, service level, suppliers, the wine list, legal permits, kitchen hygiene. Most people can sketch that much in minutes.

It is easy to define about 70 percent of a field. The devil is the last 30: the parts that are critical but do not come to mind. The grease trap the tax office asks about. The workaround everyone uses. The reason the schedule looks the way it does. The cure is unglamorous: write down what you had to learn the hard way, and let the notes accumulate until the list is complete. The last 30 percent is where a field keeps its dark context, and nobody gets it on the first pass.

A conversation has resolution the way a map has scale. At the broad level you learn the core ingredients of the dish. Zoom in and the fine grain starts to matter: quantities, seasoning, order, cooking times, how fried the steak should be, the secret in the sauce. And the finest grain never makes it onto any map at all, because the hands know it and the mouth cannot say it. That layer has its own name: tacit knowledge.

Academia solves resolution in one direction only. An academic map of a field is accurate the way a 35 by 49 metre map is accurate: every contour verified, and impossible to carry into the forest. You can spend years internalizing one, which is roughly what a degree is. Accuracy was never the problem. Access is. A good map has a low threshold and a deep conversation behind it.

Every map has a key: the small box that says what the symbols mean. Different landscapes need different keys. A forest key marks swamps, pits and density; a coastal key marks cliffs, caves and beaches. Different map makers see different things in the same terrain: a traffic planner and an urban planner map the same city at the same scale and highlight different objects. Whole schools of thought are competing keys, handing their followers different legends to the same landscape.

Here is the honest confession at the center of our method: we are not map makers. We make keys. We provide the elements that let a team map its own landscape, and we do not decide how the elements relate or what the landscape looks like. A finished, detailed map leaves no room for the reader to fill in. A key leaves the drawing to the people who actually hold the territory, and a general enough key travels: goals, roles, budget and time frame describe a website and a music festival alike.

You can tell a key is right by what happens when it is used. Overlay the maps from many conversations with many practitioners and the shared structure emerges. The best reaction a conversational map can get is a shrug: of course these are the parts. Self-evident is the goal, because it means the map matches the landscape people already navigate by.

Dark context is the unmapped part of the village: what an organization knows but never wrote down — the situated context a general AI model cannot see. The village ran on it for years, and never noticed. The stranger cannot run on it at all.

Mapping conversations is how the dark layer surfaces. Not extraction, and not surveillance: cartography, done by the people who hold the landscape, using a key that makes the drawing fast. The map is not the end of the work either. A map nobody maintains drifts gray, and a true expert holds more than the map: they know the activity on it. That part stays with people, which is exactly why the map should say where it lives.

For conversations, map making has barely begun. Computers were a professional's tool until the graphical interface arrived; most fields still hand their strangers a 35 metre chart or nothing at all. That gap is what this whole site is about.

What is a conversational map?

A taxonomy of a field: the concepts a discipline uses when it talks about its work, collected and laid out so that people can orient. It maps how practitioners actually talk, not how a textbook orders the subject, and it works as a shared interface between people with different mental models.

What is the difference between the map and the key?

The map shows how the parts of a landscape relate; the key says what the parts are, like the small box on a map that explains the symbols. A key can be provided from outside as a shared language. The map can only be drawn by people who know the territory.

How do you map a discipline?

Document its conversations. Ask practitioners how they work, what matters, and why. Keep track of the order in which things come up and note what keeps repeating, because repetition marks the core concepts. Then validate by overlaying many conversations until practitioners find the result self-evident.

Why is the last 30 percent of a field so hard to map?

Because it is critical without being memorable. The first 70 percent of a field comes to mind in minutes. The rest is the workaround, the permit, the reason behind the schedule: things people know but do not think to say. It surfaces slowly, mostly by writing down what had to be learned the hard way. That band is where dark context lives.

What makes a map of a field good?

Low threshold, deep conversation. A good map is understood in minutes and still sparks in-depth discussion. Accuracy alone is not enough: an academic treatment of a field is accurate the way a 35 metre map is accurate, and just as hard to carry into the forest.

What do conversational maps have to do with AI?

An AI is the newest stranger in the village. A general model knows every field in general and your organization not at all, and it cannot use knowledge that was never written down. A conversational map is how the village makes its landscape legible: to the new hire first, and now to the machine.