Ranking
Which to pilot first?
Enter use cases to see the ranking.
Calculator Library / AI Operations
AI Use-Case Prioritization Matrix
Score your AI ideas on six criteria with editable weights, apply gates for weak data or intolerable error risk, and get a ranked list sorted into pilot now, prepare first, and defer.
Tool
Score your AI use cases
Score = Σ weight × score ÷ 5, scaled to 0–100
Enter each candidate and score it from 1 to 5 on the six criteria, where a higher score is always better. Set the weights to match what matters to you. The tool ranks the candidates, flags any that fail a gate (data readiness or error tolerance of 1), and sorts them into tiers. Load the example to see eight candidates at an illustrative plant.
| Use case | Value weight | Volume weight | Process readiness weight | Data readiness weight | Error tolerance weight | Ease of integration weight |
|---|
Your entries stay in your browser and are saved there between visits on this device. Nothing is sent to a server. The example is made-up data.
Picture
Scores by use case
Instructions
How to use this matrix
- List your candidates, one per row. Describe the task, not the tool.
- Agree what a 1 and a 5 mean on each criterion with examples, then score together so that differences are discussed.
- Set the weights. They do not need to add to 100: the score is scaled by their total.
- Check the gates. A candidate with data readiness or error tolerance of 1 is marked “Not yet” whatever its score.
- Use the ranking to start a conversation, not to end one. A few points of difference means little. Then build a business case for the top candidates.
What This AI Use-Case Prioritization Matrix Does
The matrix ranks a list of AI use cases on six criteria: value, volume, process readiness, data readiness, error tolerance, and ease of integration. Each candidate is scored from 1 to 5, the weights turn the scores into a number from 0 to 100, and the candidates are sorted into tiers: pilot now, prepare first, or defer. A gate marks any candidate with a data readiness or error tolerance score of 1 as not yet ready.
It matches the worked example in the AI Use-Case Selection and Business Case guide. For the next step, build a case with the AI Business Case and ROI Calculator.
How the Score Works
| Element | Rule |
|---|---|
| Criteria | Value, volume, process readiness, data readiness, error tolerance, ease of integration. For every criterion a higher score is better |
| Score | Sum of (weight × score) ÷ (5 × total weight) × 100 |
| Tiers | Pilot now at or above the first threshold; prepare first at or above the second; otherwise defer. You set both thresholds |
| Gate | A score of 1 on data readiness or error tolerance gives “Not yet”, whatever the total |
| Ranking | Complete rows only (a name and six scores from 1 to 5), highest score first |
Worked Example
The pre-loaded example scores eight ideas at an illustrative 140-employee plant. With the default weights (value 25, others 15) complaint classification scores 80, invoice matching 77, and first drafts of 8D reports 75: all reach the pilot tier. Visual inspection has the highest value and volume but weak data readiness and error tolerance (2), so it scores 67 and lands in “prepare first”.
Limits and Good Practice
- A score is a way to structure a discussion. It is only as good as the inputs, so score in a group and agree what each number means with examples.
- A few points of difference between two candidates means little. Look at the criteria behind the number.
- The criteria and weights are examples. Add or change criteria in the AI Use-Case Register if your organization needs different ones.
- A high score is not an approval. Check risk, law, and data rules with the governance process in the AI Governance and Risk guide.
- Entries are saved in your browser on this device. Nothing is sent to a server.
AI Use-Case Prioritization Matrix Frequently Asked Questions
How do I score an AI use case?
Give each criterion a score from 1 to 5, where a higher score is always better, using agreed definitions with examples. For instance, a data readiness score of 5 means the data exists, is accurate and accessible, and may be used for this purpose, while 1 means it is scattered, poor, or restricted.
Why is there a gate for data readiness and error tolerance?
Because a high total can hide a deal-breaker. If the data is not usable, or a mistake would be severe and hard to catch, a use case should wait until that is fixed, however good the value looks.
Can I change the weights?
Yes. Set weights that reflect what matters to your organization. They do not have to total 100, because the score is scaled by their total. Agree the weights before you score, so that they are not tuned to favor a favorite idea.
What should I do with the top-ranked use cases?
Build a business case for each with measured baselines, check the risks, and design the human review. Then run a small pilot with success measures and a stop rule. The ranking says where to look first, not what to approve.
Is my data stored anywhere?
Your entries are saved in your own browser on this device so that they are there when you return. Nothing is sent to a server. Clear them with the Clear all button or by clearing your browser data.