What it is
A working matrix you can apply to any candidate AI opportunity. Score each on eight categories (1–5), sum, and use the result as one input into the conversation — not the decision itself.
Why it matters
Most AI backlogs are a mix of "this would be cool" and "this would save real time." A prioritization matrix surfaces the difference quickly and gives leadership a defensible first-pass ordering.
Scoring categories
- Pain. How costly or frustrating is the current state?
- Revenue/retention relevance. Is it tied to a P&L line that matters?
- Workflow frequency. How often does this happen?
- Data readiness. Is the input data accessible and usable?
- Implementation complexity. What does it take to ship a useful first version?
- Risk. Reversibility, sensitivity, brand exposure, legal exposure.
- Adoption likelihood. Will the people who need it actually use it?
- Time-to-value. How fast can it produce a result worth talking about?
Example backlog
- Weekly account-health brief for CS — high pain, high frequency, high adoption.
- AI-drafted scoping memos from discovery calls — high frequency, low risk.
- Auto-classified inbound support tickets — moderate value, depends on data quality.
- Sales rep coaching summaries from call recordings — moderate value, higher complexity.
- Auto-generated client invoices from delivery notes — high risk, low margin for error.
Next steps
Score 8–12 candidates, sort, then pressure-test the top three with the people who would actually use them. The matrix is a starting point, not the decision.
How Fascia Labs uses this
Every Diagnostic produces a scored backlog using this matrix, plus a recommended first engagement (quick win) and first strategic bet (compounding work).