# AI-Suitability Scorecard

Score any task out of 10 for how well current AI tools can help with it.

**Task:** [TASK NAME]   **Hours per week:** [#]

## How to score

Answer five questions. Each is worth 0, 1, or 2 points. Add them up.

### 1. Language-heavy
Is the work mostly reading, writing, summarizing, or reformatting words?
- **2** — Mostly words: drafts, summaries, emails, reports
- **1** — Mixed: words plus numbers, images, or software steps
- **0** — Mostly physical, interpersonal, or live

**Points: [ ]**

### 2. Pattern-based
Does the output follow the same structure every time?
- **2** — Same skeleton every time: templates, standard sections
- **1** — Loose pattern with judgment calls
- **0** — Every instance is different

**Points: [ ]**

### 3. Repeatable
How often does the task recur?
- **2** — Weekly or more
- **1** — Monthly
- **0** — Rarely, or one-off

**Points: [ ]**

### 4. Checkable
Can you tell good output from bad within five minutes?
- **2** — At a glance — I know what good looks like
- **1** — Takes real review time
- **0** — Errors could hide for weeks (tax filings, legal citations)

**Points: [ ]**

### 5. Safe inputs
Can you give an AI tool the inputs without exposing confidential data?
- **2** — Inputs are public or already anonymized
- **1** — Needs redaction, but redaction is quick
- **0** — Inputs are privileged, regulated, or impossible to scrub

**Points: [ ]**

**TOTAL: [ ] / 10**

## Reading your score

- **8–10** — High suitability. Leverage-map candidate.
- **5–7** — Medium. Possible with guardrails. Revisit after Stage 2.
- **0–4** — Low. Keep doing it yourself. That is not a failure — it may be your human edge.

## Scored examples

| Task | Score | Why |
|------|-------|-----|
| Nonprofit: draft donor acknowledgment letters | 9 | Templated, weekly, easy to check |
| Accountant: write month-end variance notes | 8 | Pattern-based; numbers still need verification |
| Teacher: adapt one lesson plan for three reading levels | 8 | Language-heavy, recurring |
| Attorney: summarize deposition transcripts | 7 | Strong fit, but inputs are privileged — score hinges on question 5 |
| Bookkeeper: reconcile bank feeds | 4 | AI miscounts; your accounting software does this better |
| Veteran in transition: salary negotiation call | 1 | Live and interpersonal — though prep for it can score high as its own task |

## Two warnings

1. A 9 does not mean AI does the task alone. You review every output. AI gets facts, citations, and numbers wrong without warning.
2. Never paste confidential data into a consumer chatbot to "test it." Stage 2 covers safe data handling.

Part of the ADAPT framework — vetslearn.com/framework