AI Governance and Use Case Approval Process
How a company decides which AI use cases go ahead: one intake form, a risk tier, the right reviews for that tier, a decision, and a register entry with an owner and a review date.
4 teams12 steps2 decisionsFree download
Free to download. Edit it in Flow Studio to change the teams and steps, ask AI for changes, or export to PowerPoint, Word or draw.io.
Step by step
The same flow as a numbered list, ready for an SOP, a BRD or a ticket.
| # | Step | Who | Next |
|---|---|---|---|
| 1 | Team wants to use AIStartRequesting team · step 2 | Requesting team | step 2 |
| 2 | Submit the intake formDocumentPurpose, data, users, vendor/model, decisions affected, human oversightRequesting team · step 3 | Requesting team | step 3 |
| 3 | Assign a risk tierLow / Medium / High using a scoring guide aligned to the EU AI ActAI governance lead · step 4 | AI governance lead | step 4 |
| 4 | High risk?DecisionAI governance lead · No: step 5, Yes: step 6 | AI governance lead | No: step 5, Yes: step 6 |
| 5 | Fast-track checksApproved tools only, no sensitive data, usage policy acknowledgedAI governance lead · step 10 | AI governance lead | step 10 |
| 6 | Privacy, security and legal reviewDPIA, vendor assessment, data residency, contract termsPrivacy, security and legal · step 7 | Privacy, security and legal | step 7 |
| 7 | Impact and bias assessmentDocumentAffected groups, error costs, subgroup testing, human-in-the-loop designPrivacy, security and legal · step 8 | Privacy, security and legal | step 8 |
| 8 | Approve?DecisionAI review board · Yes: step 10, With changes: step 9, No: step 11 | AI review board | Yes: step 10, With changes: step 9, No: step 11 |
| 9 | Change the design and resubmitRequesting team · Resubmit: step 2 | Requesting team | Resubmit: step 2 |
| 10 | Add to the AI registerDocumentOwner, tier, review date, monitoring planAI governance lead · step 12 | AI governance lead | step 12 |
| 11 | Not approvedEndAI review board · Done | AI review board | Done |
| 12 | Approved to buildEndRequesting team · Done | Requesting team | Done |
When to use it
Use it when teams are starting AI projects faster than anyone can track, and you need a light, fair way to approve them that matches regulations such as the EU AI Act and frameworks like NIST AI RMF or ISO/IEC 42001.
Who's involved
- Requesting team4
- AI governance lead4
- Privacy, security and legal2
- AI review board2
Yours works differently?
Describe your own process in a few sentences and Flow Studio draws it, with your teams and your rules.
Describe your processarrow_forwardFor the technical team
The build decisions behind this flow. Each step above also carries its technical detail.
- 01Intake captures: purpose, users, data categories (personal, special category, confidential), model and vendor, decisions it influences, human oversight.
- 02Risk tiers: Low (fast-track), Medium (privacy + security review), High (adds impact assessment, bias testing, legal, documented human oversight).
- 03Evidence pack for high risk: data sheet, eval results by subgroup, monitoring plan, incident contacts.
- 04Register fields: id, owner, tier, systems, vendor, approval date, review date, status.
How to get it right
- 1Keep the intake form to one page. A heavy form pushes people to use AI without telling anyone.
- 2Let low-risk uses through in days. Spend review time on uses that affect people's rights, money, health or jobs.
- 3Give every approved use case a named owner and a review date in a register.
- 4Re-review when the model, the data or the purpose changes, not only on the calendar.
Questions
What should an AI governance process include?
An intake for each use case, a risk classification, privacy and security reviews scaled to risk, human-oversight requirements, an approval decision, a central register, and periodic re-review and monitoring.
How do you classify AI risk?
Look at who is affected and how: uses that decide on hiring, credit, insurance, healthcare or legal outcomes are high risk; internal drafting and summarising are usually low. The EU AI Act's categories are a useful reference even outside the EU.
More process flow templates
- AI feature integrationTaking an AI feature from idea to production: pick the use case, test models against real examples, add guardrails, pass a security review, and roll out in stages with monitoring.
- RAG pipelineHow retrieval-augmented generation answers questions from your own documents: documents are split, embedded and indexed, then each question pulls the best passages and the model answers with citations.
- AI agent with tools and MCPHow an agentic AI system works through a task: it plans, calls tools through MCP servers with scoped access, asks a person before risky actions, checks results and loops until the goal is met.
- MCP server integrationConnecting an internal system to AI assistants through a Model Context Protocol server: design the tools, secure them, test them, review them and roll them out with monitoring.