When people say “user,” they usually picture someone at a desk — a knowledge worker, a laptop, an office. That picture quietly decides who a product is for, and who it isn’t. We started from a wider one on purpose: a user is anyone who touches a digital system to do their job. The operator on a factory floor. The analyst in a forensic lab. The scientist at a bench, the clinician, the logistics coordinator, the fully-remote team at a kitchen table. All users. All the same building.
A smart building of digital rooms#
We describe the product as a smart building for your team — but the building, and every room in it, is digital, not physical. A “room” is simply wherever a person touches the platform to do their work. It doesn’t matter whether they’re on a plant floor or working remotely three time zones away: the room is the same digital space, with the same feeling. That sameness is the whole ambition. Wherever the work happens, you’re touching the same layer, and it behaves the same way.
This is a big claim, so it needs a way in that isn’t “boil the ocean.” We take it sector by sector: win a signature niche in one setting, get it genuinely right, then carry the pattern outward — clinical, manufacturing, forensic, supply chain — each with its own front door but the same standard underneath. Ambition at the level of the vision; humility at the level of the entry.
Where the desk assumption actually hides#
The assumption is rarely stated, which is why it survives. It lives in defaults that nobody argued about.
It hides in the shape of the session. A product designed for a desk assumes a long, uninterrupted stretch of attention, and builds flows that punish being interrupted. Work at a bench or on a line is not shaped like that: it is short, interleaved, and frequently abandoned halfway through. A form that loses its state when you walk away has decided who it is for.
It hides in what counts as an input. Typing is the desk’s assumption. A gloved hand, a shared terminal, a device behind a screen protector, a person who needs both hands for the actual work — each of those makes typing the exception rather than the default.
It hides in the authentication model. One person, one device, one long-lived session is a desk pattern. Shared stations with fast handover are normal almost everywhere else, and a system that cannot express “this action, by this person, on a device several people use” will either record the wrong name or make people share a login. Both are worse than the problem.
And it hides in language, which is the one people notice last.
Plain words are part of the design#
If the goal is every kind of worker, the language can’t belong to just one of them. Lean on a single industry’s jargon and you quietly signal “this tool is only for them” — which works directly against the point. So we describe the product in everyday words anyone can follow, and when we show a specific setting, we keep it in language a layperson understands. Speaking plainly isn’t a marketing flourish here; it’s a design constraint that follows from who we’re building for.
It is also the cheapest constraint to state and the hardest to hold, because jargon is efficient for the people who already have it. Every specialist audience will ask for its own vocabulary, and every one of those requests is locally reasonable. Granting them produces a product that reads fluently to whoever asked most recently and slightly foreign to everyone else. So the rule is one idea, one word, chosen to be understood by someone outside the room — and specialist vocabulary lives in the content a team writes for itself, not in the surface everyone shares.
The objection worth taking seriously#
A product for everyone is usually a product for no one. That is the standard criticism of a wide definition, and it is earned: generic tools lose to specific ones in every individual sale, because the specific one already speaks the buyer’s language and models their edge cases.
The distinction we would draw is between a wide definition and a wide entry. The definition is about what the standard has to reach eventually, and it constrains architecture: it is why sessions survive interruption, why identity handles shared devices, why the vocabulary stays plain. The entry is deliberately narrow — one setting, one signature niche, done properly enough that the people in it would object if it were taken away.
The two are compatible only if going into the next sector is a configuration rather than a rebuild. That is the actual test of this position, and it is the same test as everywhere else in how we build: if reaching a new kind of worker requires forking the product, then the wide definition was a slogan and the narrow entry was the real product all along.
It is fair to say we should be judged on that rather than on the ambition. A company can say “every kind of worker” cheaply. What it costs to add the next kind is the number that means something.
How to check this#
Ask what happens to work in progress when someone walks away for ten minutes. Ask how the system records an action taken on a shared device. Ask to read a page of the product’s own interface copy and see whether it makes sense to someone outside the industry it was written for.
The widest definition of “user” isn’t a slogan; it’s the thing that decides what we build and how we talk about it. Governed digital work for everyone who touches a system to get a job done — not just the office, not just IT. If someone’s work runs through a screen, the standard should reach them.