Five formats, and how to pick one
The right one is usually decided by how much authority the room has and how much time it can give, rather than by budget.
The board briefing, ninety minutes. Inside a scheduled board meeting. One argument, the five decisions, and the questions to put to management before the next meeting. The cheapest way to find out whether the board is ready for the longer piece.
The away day or strategy day, half a day or a full day. The board away from the building, working rather than listening. Presentation first, then the work: the decision list, the stop list, the incident definition, and what the board should be shown each meeting. This is the format that produces something written.
The chaired roundtable, two to three hours. Eight to fifteen peers from different organisations, under the Chatham House Rule, on one question. Run for a membership body, a firm convening its clients, or a chair who wants to hear how other boards are answering the same question.
The private dinner, an evening. Ten or twelve chairs, chief executives or partners around one table, one question, chaired so that the conversation goes somewhere rather than circling. No pitch, no deck, no sponsor.
The leadership team session, half a day. The executive rather than the board, because the decisions the board wants made are usually made a layer below it.
What I am doing in the room
Chairing, mostly. The material is the easy part and it is already written: the research behind every session is public at the research, and a board can read it before the day rather than be told it on the day.
What a room like this needs is somebody who will keep it on the decision rather than the technology, who will name the thing everybody is avoiding, and who has no stake in which way it goes. I sell no software, take no vendor’s money and build nothing, which is the whole reason a chair can let the conversation be honest in front of me.
Sessions run under the Chatham House Rule where the room wants it, which is most of the time and always at a dinner or a cross-company roundtable.
What the board leaves with
Written, and short enough to be read. Four things, which are the same four whatever the format and the time available changes only how far through them the room gets.
The list of decisions machines may already make in the company’s name, and who approved each. The list of who can stop each system, how long a stop takes, and when it was last rehearsed. A definition of an AI incident, written before there is one. And what management should put in front of the board each meeting, which is the thing that makes the other three survive past the away day.
The argument underneath all of it is at board oversight of AI, including what Provision 29 of the UK Corporate Governance Code now asks of a board and what it does not.
Preparation, because the day is mostly won before it
A fact-find with the chair or the company secretary, a read of whatever management has already produced, and where the board wants it, a short anonymous survey of the executive before the day so the room starts from its own evidence rather than from mine. That survey has changed the shape of more sessions than any other piece of preparation: it is usually the first time a board sees, in its own people’s words, which judgement has already been handed over.
Sessions are built for the room. The same argument, at the altitude the room operates at, with examples from the sector it is in.
Where this sits next to the keynote
A keynote is the right format when the job is to change how a large room thinks, often often the thing that happens first. The formats on this page are what happens when a smaller room has to decide something. If what you want is the stage version, that is at keynotes, and it costs less.
If what you want is a written independent view of a strategy you have already built, without a room at all, that is a second opinion on your AI strategy.
