Part 2 — Weigh Them
The Value × Feasibility Matrix
A use case doesn’t earn a green light for being exciting. It earns one for scoring well on two axes at once.
A list of candidates only helps if you can compare them on the same terms. Score each one on two axes — value and feasibility — independently. A use case doesn’t earn a green light for being exciting; it earns one for scoring well on both, at the same time.
Value: is it worth the effort at all
| Dimension | What high value looks like | The red flag |
|---|---|---|
| Frequency × duration | Done often enough, and takes long enough, that fixing it frees real hours, not a task run twice a year | A task so rare that setup cost outlasts the benefit |
| Cost of the status quo | A clear cost in staff hours, error rate, rework, or lost business from doing it the current way | “It’s just annoying.” Annoyance isn’t a number. |
| Revenue or risk exposure | Fixing it opens revenue (faster quoting, more follow-through) or closes a risk (compliance, missed renewals) | No real connection to either — it just feels like it should matter |
| Visibility | Solving it produces a number leadership will recognize in 90 days | Only the person doing the task would ever notice it changed |
Feasibility: can this business actually pull it off
| Dimension | What high feasibility looks like | The red flag |
|---|---|---|
| Data readiness | The data this depends on already exists somewhere structured | The data would have to be built before the AI could touch it |
| Task structure | Rule-based and repeatable: the same decision, made the same way, most of the time | Judgment-heavy, hinging on context or expertise an AI can’t see |
| Integration load | Lives inside one system, or close to it | Solving it means touching four tools and someone’s inbox |
| Change resistance | The people who’d have to change behavior already want this fixed | Whoever owns the current process is attached to it |
Score each candidate 1–5 on each axis and record the result:
| Candidate | Value (1–5) | Feasibility (1–5) | Quadrant | Decision |
|---|---|---|---|---|
| e.g. Invoicing follow-up | 4 | 5 | High-value / high-feasibility | Take to Screen It |
Score each candidate 1–5 on value, 1–5 on feasibility, and plot the results. The quadrant that matters for a first use case is high-value, high-feasibility — not the most ambitious one on the list, the most winnable one. High-value, low-feasibility candidates aren’t wasted; they go on a “later” list, and showing that list to the person funding this is often what earns you the budget for the next one.
The use case you pick here is the one that goes to Screen It next — this matrix tells you which candidate is worth building a real Value, Cost & Risk Canvas for. Skipping straight to Screen It on the first idea that walked in the door is exactly the mistake this step exists to prevent.