BNI Afterwork presentation

BNI Internationals hosted an afterwork in Stockholm, put together by art dealer Alex Colard and the BNI Internationals team. People came from several different BNI chapters, guests were welcome, and the whole thing ran in English with a fair amount of Swedish in the room anyway.

I was the speaker of the month. Seven minutes, then questions.

What follows is the whole talk, more or less as I gave it.

First, what BNI actually is

If you have not come across it, BNI stands for Business Network International. It is a referral organisation. Chapters meet regularly, and the point of the meeting is that members pass business to each other rather than compete for it. Their phrase for this is Givers Gain.

The structural rule that makes it work is one seat per profession per chapter. One accountant, one lawyer, one printer. If you are in the room, nobody sitting next to you is trying to take your work, so recommending each other costs you nothing.

That is worth understanding before you present to a room like this, because it changes what a talk is for. A BNI room is not a room of prospects. It is a room of people who will repeat what you said to somebody else next week. So the job is not to pitch. The job is to be repeatable.

This particular event was an afterwork rather than a chapter meeting. Everybody got thirty seconds to introduce themselves, the ticket covered a drink and pizza, and registration went through the BNI Internationals team.

What we actually run

I opened with the stack, because it is the most immediately useful thing I can hand a room of operators and it costs me nothing to give away.

What for	What we use
Sending	Salesforge, somewhere between a thousand and ten thousand emails and LinkedIn messages a week depending on the client
Finding people	LinkedIn Sales Navigator and Apollo
Video and ads	Gemini and Higgsfield, and that one is still a work in progress
Hosting and deploys	UpCloud and Coolify
Running the operation	Claude, plus our own harness on top of it

None of that is a secret and most of it is not expensive. Which sets up the actual point.

The problem is not the tools

I asked for a show of hands. Who has tried to use AI for real work this year, not messing about. Most hands went up. Then: keep them up if you feel like you are getting anywhere near what it can actually do.

Almost every hand went down. That happens in every room.

Here is what I think is going on, and it has nothing to do with which model you picked.

Picture somebody who has been building their company for years. Everything about how that business works is in their head. Then a new person joins. Could be an employee, could be a contractor, could be an AI agent, it genuinely does not matter which.

That founder cannot hand over years. There is no meeting long enough. So the new person gets crumbs, and then everybody is surprised when the output is generic.

You would never do that to a person and expect a good result. We do it to AI constantly.

So the work is getting as much out of that head as possible, in a form both a human and a machine can actually read. There are two ways in, and I brought an example of each.

Example one: they already had the material

The first is a fellowship programme at a US university. New intake every year, and a new fellow puts in around ten hours a week.

They had plenty written down. Four different formats, in fact: a handbook, a set of decks and documents, live sessions, and recordings that fellows are expected to work through. The problem was not absence. It was that getting through all of it took the best part of a month, and most of it did not stick.

What we did was run every one of those formats through our agent and pull out what actually mattered, into a playbook the agent can answer from.

So now a new fellow does not start with a month of material. They start by asking. What am I responsible for. What should I be doing this month. What should I read first.

The part I care about most is that it meets them where they are. Everybody does not get the same firehose. You only ever learn a new thing through something you already know, and an agent that knows what you have already covered can pitch the next thing correctly. A PDF cannot do that.

And when it does not know something, it does not invent an answer. It goes to whoever does know, gets it, and keeps it. So next year's intake starts where this year's finished instead of starting over.

Example two: he had nothing

The second example was sitting in the room, which made it easy.

He runs a referral network here in Stockholm. People, who they know, and making introductions between them. When we first spoke he told me he had a document for this and a document for that, which I want to defend rather than mock, because that is what building a company actually looks like. You are not being disorganised. You are surviving. Nobody sits down and structures the thing while they are still trying to make it work.

So he had nothing to hand in. The opposite starting point from the fellowship programme.

What we did instead was point the agent at him and let it interview him. It asked the questions, he answered, and the playbook came out of the conversation rather than out of a folder.

His day to day use is tracking the people he meets. He takes a screenshot of somebody's LinkedIn, sends it to the agent, says where he met them and why he wants to follow up, and it handles the reminder. That used to live in a spreadsheet.

The more interesting part is what happens next. His company is going to grow, and at some point he brings somebody in. Normally that person costs him a week of his own time just explaining who is in the network and why any of it matters. Now they get handed something that already knows.

How it is actually built

Two layers, and this is the whole product.

The company playbook. How this place works and where everything lives. Written once, read by everyone. It describes the seat rather than the person.

A personal playbook, one per person. Their role, what they own, what they have been doing. Fifty people means fifty personal playbooks, all sharing the one company playbook underneath.

The bit that gets the reaction is what happens when somebody leaves. The personal playbook stays. You hand it to whoever takes the seat, and the next person starts with what the last one knew instead of with nothing.

That is offboarding, and almost nobody plans for it. Everyone treats knowledge as something that arrives with a hire and evaporates when they resign.

The thing to take home

One line, and it is the only thing I actually needed anybody to remember:

You cannot delegate what you have not written down. Not to a person, and not to an AI.

Every tool in that table at the top is fine. They are all good. They are also all useless if the only place your company exists is inside your own head.

---

Thanks to Alex Colard and the BNI Internationals team for putting the evening together, and to everyone who came and asked something afterwards. If you were in the room and something here was useful, or if I rushed past the thing you actually wanted to hear, tell me: mo@opmore.io.