The value case with the non-AI alternative ruled out

In short. A use case enters the portfolio when it carries a value case that includes the non-AI alternative, assessed and ruled out in writing. Without that step the comparison runs between doing the project and doing nothing, which is a comparison that always wins. With that step it becomes clear how often the problem is solved by a process change, a query or a mandatory field in a form.

It is the cheapest control of all and the one skipped most often, because it asks people to question a project somebody has already announced in a meeting, and because the honest answer sometimes leads to doing nothing.

What a value case that holds contains

Element	What has to be there
Measured problem	The starting number, not the manager's perception
Baseline	How the process performs today, over a defined period
Non-AI alternative	The simplest option that would solve the same problem
AI option	What gets built, on which data, with what integration
Total cost	Licences, integration, maintenance, training, oversight
Risk and classification	Where the system lands and what that entails
Stop criteria	The conditions under which the project ends

The alternative row is what changes the conversation. It forces whoever proposes to face the question the board will ask anyway, and lets them arrive prepared rather than defensive.

Why the non-AI alternative has to be written

In the projects we assess, a sizeable share of proposed use cases has a cheaper alternative nobody had put on the table. A slow approval flow is often fixed by changing the authorisation threshold, ineffective document search improves by reorganising folders and enforcing a few mandatory fields, and a load of repetitive requests drops when the page generating the question is corrected rather than when an assistant is built to answer it.

Writing the alternative doesn't mean choosing it. It means stating why it was ruled out, and that sentence is worth more than ten slides on the potential of the technology, because it shows the decision was taken rather than inherited.

When no alternative exists, the value case comes out stronger. A use case still standing after being compared with the simplest option is a defensible use case.

Total cost, with the lines that are always missing

Initial estimates contain licences and development days. What is almost always missing is integration with existing systems, the time of the people validating outputs in the first months, when trust has to be built, the oversight after release, and the cost of exit if the project has to be stopped halfway.

We have written about this in detail in what AI projects really cost and in the AI cost nobody budgeted for. The practical rule is that the most expensive forgotten line concerns people, not technology.

Who signs

A value case without a name at the bottom is a study document. The signature establishes who answers for the starting number and who will decide, on the agreed date, whether the project continues.

That same person has to hold the authority to stop it, because if whoever signs cannot close the project the stop criteria stay decorative and the initiative runs on inertia until the budget is gone. It happens more often than people admit.

Where to start

The method for selecting use cases is in how to choose the first use case and in prioritising AI use cases.

The decision criteria are described in go/no-go criteria, while the six-dimension assessment framework is in AI Canvas.

To build the value case for your own use cases you can book an assessment session.