The deck is the problem.

Stalled AI initiatives nearly always have a forty-page strategy deck attached — twenty initiatives, a 2x2, six recommended, sponsors named and owners not. Twelve months on, nothing is in production, and the deck is the reason: too long for anyone who could kill an initiative to read, too vague for anyone who could ship one to act on. The fix is to throw it out and replace it with one page per initiative — a verdict, an owner, and a date — that an executive can read end to end and a builder can act on without another meeting.

It always arrives the same way. A company gets serious about AI. A senior person, often new in seat, commissions a strategy, and it comes back as forty pages: twenty initiatives scored on a 2x2, a recommended portfolio of six, sponsors named. No owners. Twelve months later, two of the six are still alive in the sense that someone is listed against them. None are in production. The deck has become the thing everyone points to when asked what the company is doing about AI — which is exactly why nothing is happening.

The deck is not a symptom of the problem. It is the problem. It is too long to be read by anyone senior enough to kill an initiative, and too vague to be acted on by anyone junior enough to build one. It exists to be referenced, not executed against. Every month it stays the authoritative document is a month no one has to decide anything.

The replacement is one page per initiative, in plain English: a verdict, an owner, a date. Short enough that an executive reads it end to end. Concrete enough that a builder acts on it without another meeting. If you have the forty-page deck, you do not need another strategy session. You need someone willing to write the one-pagers and put a verdict on each. That is most of what we do.