Pillar 04 · Accountability
C2 RACI for Each AI Use Case
A wrong answer should be someone's job.
Form Workbook
What it's for
An AI use case combines sources, definitions, a model, prompts and sometimes an agent. When an answer is wrong, the data team says the data was fine, the AI team says the model did what it was asked, and the business says it doesn't own the tool. A RACI agreed before go-live settles who does what across the whole life of the use case, from defining it to retiring it.
What you get
- RACI template Eighteen activities, ten roles, one accountable person per activity.
- Example EX-04 A completed RACI for the investor commentary drafter, a High-tier agent.
Worked out for you A count, Check
How it connects
One RACI per S1 use case, agreed before the go-live gate; C1 appoints the roles it names, and C3 points to it for each agent.
A flavour of the worked example
| Activity | 01 · Define the purpose, users and what a good answer looks like |
|---|---|
| Use-case owner | A/R |
| Agent owner | C |
| Risk, legal and compliance | C |
| Check worked out | OK |
A few fields from one record of the toolkit's worked example, with long text shortened.
Next step
Guided build coming next.
Request access
Ask for access to the guided builds, including this kit's when it is ready.