Minimum Viable AI Checklist
Use this checklist to determine whether a use case is suitable to be implemented as a Minimum Viable AI — a low-risk, high-learn validation of AI in action.
Assign 1 point for each "Yes".
✓ 18–20 points: Excellent candidate for MVAI
⚠ 15–17 points: Good candidate, refine the scope or data
❌ < 15 points: Reconsider or reframe before proceeding
1. Business Relevance
| S.No. | Question | Yes/No | Comments |
|---|---|---|---|
| 1. | Is the problem or opportunity clearly defined in business terms? | ||
| 2. | Does solving this bring measurable business value (e.g., cost, time, growth)? | ||
| 3. | Is there a stakeholder or business unit owner invested in the outcome? | ||
| 4. | Is there an existing workflow or decision this can support or enhance? |
2. Data Feasibility
| S.No. | Question | Yes/No | Comments |
|---|---|---|---|
| 1. | Do we have access to relevant data (even if not perfect)? | ||
| 2. | Is the volume and variety of data sufficient to test a simple model? | ||
| 3. | Are there basic features/inputs already available without major engineering? | ||
| 4. | Are data privacy, ethics, and compliance considerations manageable at this stage? |
3. Technical Simplicity
| S.No. | Question | Yes/No | Comments |
|---|---|---|---|
| 1. | Can we start with a simple model or heuristic (e.g., rule-based, logistic regression)? | ||
| 2. | Can we build this with existing tools, platforms, or cloud services? | ||
| 3. | Is the use case narrow in scope, with clear boundaries? | ||
| 4. | Is explainability important (and achievable) in the initial phase? |
4. Pilot Readiness
| S.No. | Question | Yes/No | Comments |
|---|---|---|---|
| 1. | Can we test this in a sandbox or limited environment (e.g., one region/team)? | ||
| 2. | Can we measure pre- and post-pilot impact (baseline defined)? | ||
| 3. | Are user feedback loops and improvement cycles planned? | ||
| 4. | Can the pilot run for 6–12 weeks with current resourcing? |
5. Adoption & Ownership
| S.No. | Question | Yes/No | Comments |
|---|---|---|---|
| 1. | Do we know who will use or benefit from the AI outcome? | ||
| 2. | Has change enablement (communication, training) been considered? | ||
| 3. | Is there a plan for ownership transfer or scale if the pilot succeeds? | ||
| 4. | Is the MVAI aligned with the team’s strategic priorities? |
If you can define the problem, access some data, and validate an outcome in weeks — you’re ready for an MVAI.
