The AI agenda your budget holder approves
Whoever releases the budget does not approve enthusiasm or a list of tools. They approve a sequence with criteria: what comes first, what it costs, what it proves, and what makes it stop.
Turn several loose AI ideas into a one-page plan, with a defined order, a cost-and-return sum for each one, and a written rule for when to continue and when to stop.
Sign up once to unlock this course
Course 01 is open. For courses 02 to 05 we ask for six fields. It is a single signup: done once, valid for the whole track. It is not a free diagnosis and it does not trigger automatic sales contact.
A list of ideas is not a plan
What usually reaches the decision table is a list: ten opportunities, a paragraph each, all marked as priorities. No order, no cost, and no statement of what each one would prove. The meeting ends asking for more study, and the year goes by.
A plan differs from a list in three things: the order has a justification, each item has a value attached, and it is written down what happens if the first one does not work.
The first initiative is the one that pays for the others
The first one should not be the one with the highest potential. It should be the one that produces the clearest proof while depending on the fewest things. Its job is different from the others: it has to prove, inside your company, that this leaves the slide deck and shows up in the results.
Five criteria for choosing, which survive the first hard question: the current cost is already known; the volume is already measured; the data needed already exists today; the area has someone willing to own it; and the result is something finance knows how to check.
One page per initiative, not thirty
A justification nobody reads protects no decision at all. The version that works fits on one page and answers six questions: what work changes; what it costs today; what it will cost; what has to exist for it to work; which risk the company accepts; and how we will measure it in ninety days.
The current cost of each case multiplied by the number of cases per month. It is the only number that does not depend on an optimistic projection, because the company already pays it today.
Every initiative needs it written down when to stop
The question that separates a mature plan from an optimistic one is simple: what makes this stop? Without that answer in writing, weak initiatives consume budget through inertia, because nobody wants to be the person who cancelled.
Four legitimate reasons to stop: the cost per case went past the limit; the share of exceptions stayed above forecast for two reviews in a row; a dependency was not resolved by the deadline; or the surrounding process changed and broke the original premise.
Who reviews it, and how often
A plan with no review date becomes a filed document. The minimum that works is a review every quarter, with three facts per initiative: where it stands, how much it has spent, and what it has proved. Anything that has not moved in two reviews goes back to the queue or is closed.
The five questions you need to be able to answer
1. Why this initiative first and not another?
2. What does it cost and where does the money come from?
3. Who answers for the result?
4. What do we do if it does not work?
5. How will we know it worked, with which number and in how long?
A one-page plan that answers those five gets approved far more often than a thirty-page document that answers none.
Basis
The plan described here is the same way any investment decision is defended in a company, with a board or without one: order, value, owner and stopping condition. AI is no exception.
When the agenda becomes an investment decision
If your first initiative already has a known cost and a proof criterion, there is a basis to assess execution. Prumo Discovery organises three things before any money is committed: a measure of the current situation, whether it is feasible, and a recommendation.