# Write it down instead of meeting

8 February 2026 · Essay · 7 min read

By [Ahsan Fazal](https://axiomatic.digital/en/about#ahsan), Founder and CEO

For the people who build, a meeting is the opposite of work. My rule: don’t meet if you can write down what you wanted to say. We decide on paper, with a date and a name, before we build, and we write for the reader’s decision, not for our own overview.

I do most of my work in the evening, early in the morning and at the weekend, as now, on a Sunday afternoon. Not because I cannot stop, but because those are the only hours without meetings. So they are the hours in which I get the most done.

This afternoon I wrote down that meetings are, for me, literally the opposite of work. Below it I wrote the only rule I could think of, in English, the way I thought it: “don’t meet if you can write what you’d like to communicate in the meeting as structured document”.

And I asked: is that enough?

This piece is my attempt to answer that question myself.

## What a meeting costs

Anyone who builds needs long stretches of time in which nobody asks anything. A fault in the way Inora retrieves a source is not fixed in the twenty minutes between two appointments. You need the whole problem in your head, and that takes an hour before you do anything with it at all. One meeting in the middle of the morning breaks that morning into two halves, both too short.

A meeting also costs everyone the same hour, at the same moment, at the pace of whoever talks longest. A document costs the writer an hour and each reader a few minutes, at a moment the reader chooses. A reader who already knows skips a paragraph. One who missed something reads it twice.

## What writing forces

The difference is not only time. In a meeting you can think out loud on other people’s time. Half a thought sounds whole in a room, and nobody sees the gap until someone starts building on it the next day. On paper you see the gap yourself, before anyone else does. Writing forces you to finish thinking before you ask anything of anyone.

A text also has a date and a name. After a meeting everyone has their own memory, and three weeks later those are three different memories. After a document there is one text. People forget; a text does not. If you claim something, you must be able to show it. And if you decide something, you answer for it, with your name underneath.

That is why we decide on paper, with the date and the name of the person who wrote it, before we build. On 13 June 2025 I shared the eight-week plan for Inora: the foundation first, the features after. An environment we control ourselves, the source with every answer, and tied to no single language model. On 4 July I wrote the first architecture: each organisation a closed space of its own, each document with its structure and its authority, and with every question the role of the person asking. Both were in place well before the first care organisation started working with Inora on 4 September.

Anyone who wants to know today why Inora is built the way it is does not have to ask anyone. It is written down, with the date.

## A text can be shot down

A document has one more advantage, and I only saw it later: you can shoot it down. On 25 November I wrote down [the grammar](https://axiomatic.digital/en/updates/a-grammar-for-how-knowledge-connects) of how knowledge connects in Inora: twelve kinds of link in four families, their properties and the rules that follow from them. I added that it was a working document. I set out claims as if they were true, only to get the ball rolling, and asked for every wrong one to be shot down without mercy.

A text does that, and a meeting does not. A claim on paper stands still. You can strike it through, write down why, and everyone sees the correction. A claim in a meeting is gone before anyone has been able to test it. Or it lingers, and nobody remembers that it was wrong.

## Who are you writing for?

A document is not yet a good document. Twice in five weeks I sent the wrong one first.

On 20 October I was asked to check and improve a budget. At 17:33 I sent my version, ten pages: how everything fitted together, phase by phase. But the reader had to decide on line items, not on how things fit together. Six minutes later I wrote that I had made the classic mistake. A technical mind wants to explain how everything fits together. The person who has to decide wants the information divided by the kind of decision. At 17:54 I sent two pages, with only the lines we were actually going to carry out.

On 25 November, the same evening as the grammar, it went the same way. At 18:21 I first sent my own summary, four pages in mathematical notation. I already knew as I pressed send: it was a summary for myself, not communication. Five minutes later I asked for it to be ignored: it was noise. At 18:43 the same content came in plain language, also on four pages. That was the signal.

My overview is for me. The reader has a decision to make, and a good document is divided by that decision, not by how it sits in my head. That is what “structured” means in my rule. Not tidy, but divided by what the reader has to do with it.

The picture I have of Inora is not automatically anyone else’s. How the maker sees a product and how a board member or a nurse sees it drift apart until someone writes it down for their question.

## Three readers, three decisions

When it goes right, it looks like this.

A care organisation had asked us questions about privacy and responsible AI. The person who handles privacy in a care organisation has a say in whether staff may work with Inora. On 1 October [I answered the questions](https://axiomatic.digital/en/updates/our-answers-on-privacy) on paper, on seven pages with my name underneath: no training on conversations, no name or email address attached to a message, hosting in Europe, and a knowledge library that grows only after experts have checked what goes in, never by itself. Anyone who wants to know in a year’s time what we promised then can read it.

On 18 December the steering group of South Limburg’s care network met: the board members of the participating organisations. That meeting has its place, because that is where they take the decisions about Inora. But board members do not need an architecture. I set out [our picture of Inora](https://axiomatic.digital/en/updates/our-picture-for-the-steering-group) for them on paper, in one table and one illustration: the knowledge of the profession for everyone, each organisation’s own knowledge on top, every gap back to the person who keeps the knowledge, and the professional stays responsible. The steering group took it as its own picture.

Since 21 January every organisation that works with Inora has received [its own report](https://axiomatic.digital/en/updates/per-team-never-per-employee), written by Mustehsan: how many people use Inora and come back, what for, which answers they rated good and which topics have no good answer yet, each with a count. Never who asked what. The report is divided by what the organisation has to do with it: which knowledge to add first.

Three readers, three decisions, three documents. None of the three is our overview of Inora. That overview exists too, in the plan and the architecture. But it is for us.

## Is it enough?

Back to this afternoon’s question. No, the rule is not enough. It misses two things.

The first: not everything can be written down, and that part matters most to us. What a care worker knows after years on the ward is often in no protocol at all. You do not collect it with a document. You hear it by asking, preferably where the work is done. That is exactly the knowledge that has to go into Inora. I do not skimp on that conversation.

The second: a document decides nothing. Sometimes a real difference of view remains after the reading. Then we talk, briefly, with the text on the table. And then the conversation ends in a new text: what was decided, by whom, on what date.

So the rule gets sharper. Don’t meet to say what you can write down. Talk to hear what is not written anywhere yet. And then write down what you decided.

This piece is a text too. It can be shot down.

## Do you work in care and use Inora?

Sign-in and help go through your own organisation: your team’s project lead and ambassadors can help you.

## Read on

- [Start with the work, not the technology](https://axiomatic.digital/en/updates/start-with-the-work.md): Inora’s first architecture starts with the work of care, not with a language model. It fixes three things: a closed space for each organisation, every document with its structure and authority, and, with every question, the role of the person asking it.
- [Axiomatic B.V. founded in Delft](https://axiomatic.digital/en/updates/axiomatic-founded-in-delft.md): Axiomatic B.V. has been founded in Delft, with Inora as its first product. As managing director and statutory director I answer for what we make and for the decisions behind it.
- [Our picture of Inora, for the steering group](https://axiomatic.digital/en/updates/our-picture-for-the-steering-group.md): One table and one illustration for South Limburg’s board members. The profession’s knowledge for everyone, each organisation’s own on top, every gap back to the people who keep that knowledge. The steering group adopted it on 18 December 2025. For 2026 we commit to each organisation’s own knowledge.
