Enterprise AI Investment Decision Brief
A framework for putting an AI investment decision to leadership: the decision asked for, the options, what it takes, how it can fail and how you will know.
Free to download. Complete it on screen or print it.
What the brief asks
- The decision requested, in a sentence a reader could answer yes or no to, and why now
- The change in how work is done, described before the technology
- Options considered, including doing nothing and a smaller version, in a table of what each would take, its main risk and why it was chosen or set aside
- What it takes: people, data, approvals and process change, with a named owner
- How it can go wrong, how you will know it is working, and what would make you stop
- What you need from leadership, the named owner after launch and the date of the next review
Writing it
- 01
Answer the rule for every section
For each entry, state what you know, what you are assuming, and what would change your mind.
- 02
Attach the analysis, do not paste it
Where an entry needs a figure, write the reasoning in words and attach the supporting analysis separately, with its source.
- 03
Write for a reader who was not in the project
The brief is meant to be read in a few minutes. Cut anything that needs the project history to follow.
What it does not replace
A brief supports a decision. It does not replace the analysis behind it, and it should link or attach that analysis with its source. To rank candidate use cases before you write it, use the AI Use-Case Prioritization Tool and Matrix.
On this page
How the options table gets filled
This is where a decision brief usually fails: the options are the favored one and two that were never considered. Fill the table like this.
| Option | What it would take | Main risk | Why chosen or set aside |
|---|---|---|---|
| Do nothing | No spend, and an agreed account of what the current way costs | The cost of waiting is not seen until later | Set aside only if the cost of waiting can be shown in words |
| A smaller version | A single team and process, and a first step that can be reversed | May not show the effect | Chosen when it can be stopped cheaply |
| The proposal | The people, data, approvals and process change listed under what it takes | Named, with who is affected and how soon anyone would notice | Chosen because the smaller version cannot test the central assumption |
A leader asked to approve a brief with no stop condition is being asked to approve an open commitment.
What do people ask before writing an AI investment decision brief?
What should an AI investment decision brief open with?
The decision requested, in a sentence a reader could answer yes or no to, followed by why now. Next comes the change in how work is done, described before the technology. A leader then knows what is being approved, by whom and what has changed.
Which options belong in the options table of the brief?
Doing nothing, a smaller version and the proposal itself. Each row says what it would take, its main risk and why it was chosen or set aside. Briefs usually fail here when the options are the favored one plus alternatives that were never really considered.
Where do the figures go in a decision brief?
Write the reasoning in words and attach the supporting analysis separately, with its source. The brief is meant to be read quickly by people who were not in the project, so detail and history belong in the attachment, and anything needing the project history to follow is cut.
Why does a decision brief need a stop condition?
Because a leader asked to approve work with no stop condition is approving an open commitment. Agree the conditions in advance with the sponsor and the finance partner, such as a quality level, a cost that triggers review, or a risk event that pauses the work.
Framing a real investment decision?
Describe what you are weighing in a few lines. A senior practitioner replies with a suggested first conversation.