AI Use-Case Scorecard
Score each AI idea on value and readiness to see which to start now, which need prep work first, and what review each needs before launch.
Ideas
Example scores are illustrative. Select an idea to edit it, or remove the examples and add your own.
Score the idea
Value and readiness (1 is low, 5 is high)
Risk questions
Where it starts
Your summary
Sorted with quick wins first. Keep a copy for your team or your next planning meeting.
Want a second opinion? Get a free review.
Send your scorecard to the Surfstack team and we’ll reply with:
- Which quick win we’d pilot first, and how
- What it would take to close the readiness gap on your strategic bets
- What the governance review should cover for any Tier 3 or Tier 4 ideas
Optional. Only your summary and the details you type here are sent. No mailing list.
How the scorecard works
Every idea gets two numbers. Value is how much time, cost, or customer benefit it changes. Readiness is the average of three scores: whether it can be built with tools you have, whether the data is ready, and whether people will actually use it. Four risk questions then set the review it needs before launch.
The four scores
- Value
- How much time, cost, or customer benefit does this change? 1 means little, 5 means large.
- Feasibility
- Can it be built with tools you have or can approve? 1 means hard, 5 means easy.
- Data readiness
- Is the data accessible, current, and permissioned? 1 means not ready, 5 means ready.
- Adoption readiness
- Will the people who must use it change how they work? 1 means resistant, 5 means eager.
Where each idea lands
High value is 4 or 5. High readiness is an average of 3.5 or more.
Quick win
High value, high readiness
Start here: pilot on the current platform with a test set and a human review gate.
Strategic bet
High value, low readiness
High payoff but not ready yet: close the lowest readiness gap, then build.
Nice to have
Low value, high readiness
Low payoff and easy to do: offer as self-service within approved tools.
Park
Low value, low readiness
Low payoff and not ready: revisit at the quarterly review.
Risk tiers and what each needs before launch
The highest tier that applies wins. Any automated decision about people, accounts, or contracts is Tier 4. Customer-facing output or customer, personal, or financial data is Tier 3. Internal systems or documents alone is Tier 2. Everything else is Tier 1.
| Tier | Applies to | Required before launch |
|---|---|---|
| 1: Low | Staff productivity, no sensitive data, output stays with the user. | Approved tool; Acceptable-use training |
| 2: Moderate | Internal systems and documents, staff-only output. | Owner sign-off; Test set; Access-trimmed retrieval |
| 3: High | Customer-facing output, or customer, personal, or financial data. | Governance review; Named human approver on every output; Monitoring |
| 4: Restricted | Automated decisions about customers, staff, accounts, or contracts. | Not permitted without executive and governance approval |
Common questions
How do I prioritize AI use cases?
Score each idea on value and on readiness, where readiness is the average of feasibility, data readiness, and adoption readiness. Ideas with high value (4 or 5) and high readiness (3.5 or more) are quick wins to pilot first. High-value ideas that aren't ready yet are strategic bets: fix the weakest readiness score before building.
What review does an AI project need before launch?
It depends on risk. Staff-only tools with no sensitive data need an approved tool and acceptable-use training. Anything customers see, or that touches personal or financial data, needs a governance review, a named human approver, and monitoring. Automated decisions about people, accounts, or contracts should not proceed without executive and governance approval.
Is my data sent anywhere?
No. The scorecard runs entirely in your browser and nothing you enter is stored or sent to Surfstack. If you choose to request a free review, only the summary and the contact details you type into that form are emailed to our team.
Can I use this for a public-sector agency?
Yes. The risk questions cover resident-facing output and automated eligibility decisions, which are common concerns for municipalities and utilities. Treat the tiers as defaults and swap in your own AI governance policy where it differs.