Every nominated process runs through these four weighted dimensions. Scores are 1–5 with concrete behavioral anchors. Weighted composite determines pipeline tier.
Effort and Risk are inverted — higher scores mean easier implementation and lower risk, so all four dimensions point the same direction (higher = better candidate).
Captures direct labor savings, error-cost reduction, and speed-to-revenue uplift. Use whichever metric is most defensible — labor hours, hard-error cost, or time-to-value. Where multiple apply, take the strongest signal; don't double-count.
| Score | Labor Savings | Error Cost Reduction | Speed-to-Revenue |
|---|---|---|---|
| 5 | >5 FTE or >$500K/yr saved | Errors eliminated carry >$200K/yr hard cost or regulatory penalty | Unlocks material new revenue (>$1M/yr) by removing a bottleneck |
| 4 | 2–5 FTE or $100K–$500K/yr | $50K–$200K/yr in error costs eliminated | Accelerates revenue recognition by weeks, or enables $250K–$1M new revenue |
| 3 | 0.5–2 FTE or $25K–$100K/yr | $10K–$50K/yr in error costs eliminated | Noticeable cycle-time improvement but no direct revenue unlock |
| 2 | 0.1–0.5 FTE or $5K–$25K/yr | $2K–$10K/yr in error costs eliminated | Minor speed gain, hard to attribute revenue impact |
| 1 | <0.1 FTE or <$5K/yr | <$2K/yr in error costs or errors are not tracked | No measurable revenue or cycle-time effect |
* FTE = full-time equivalent (40 hrs/wk). Use fully-loaded cost (~$80K–$120K/yr per FTE as a default if actual is unknown). Hybrid scoring: if a process saves 3 FTE but the labor is offshore at $25K/FTE, score conservatively (the $ anchor dominates).
Measures whether the process connects to a stated company priority — not whether it "feels important." Requires a named company goal or initiative (from annual plan, OKRs, board priorities). If you can't name the priority it aligns to, it's a 1 or 2.
| Score | Anchor | Test |
|---|---|---|
| 5 | Mission-critical enabler. A top-3 company priority cannot be achieved without this automation — or the automation is itself the priority. | Would the executive sponsor of that priority personally care about this project's status? Can you write the one-line connection without stretching? |
| 4 | Directly enables a top priority. The automation removes a known constraint or unlocks speed/capacity for a named strategic initiative. | Is the priority named in QBRs, board decks, or all-hands? Does the automation appear in the initiative's dependency map? |
| 3 | Supports a priority area. The process lives inside a function or workstream that is strategically important, but the automation is one of several enablers — not the lynchpin. | Would removing the automation delay but not derail the priority? |
| 2 | Tangential to a priority. The process touches a priority area indirectly, or the connection requires a multi-step "if-then" chain to articulate. | Can you describe the link in one sentence that a busy VP would nod at, or does it take a paragraph? |
| 1 | No connection to any stated company priority. The process may be important operationally but isn't linked to a strategic goal. | If the automation never shipped, would any strategic KPI move? If the answer is "no," score 1. |
Inverted scale — higher score means easier to implement. Evaluates systems integration complexity, data quality, and estimated delivery timeline. If any single sub-factor is a showstopper (no digital access at all, data is pure narrative text with no pattern), cap at 2 regardless of other factors.
| Score | Systems Integration | Data Quality | Estimated Timeline |
|---|---|---|---|
| 5 | Single modern system with well-documented REST API; or fully self-contained (Excel → Excel) | Structured, validated, clean — fields are predictable, formats consistent, no cleansing needed | <4 weeks from kickoff to production |
| 4 | 1–2 systems, APIs available but may need some middleware or orchestration | Mostly structured, occasional inconsistencies that can be handled with simple rules | 1–2 months |
| 3 | 2+ systems, mix of API and file-based integration (SFTP, CSV export), or a legacy system with a workable connector | Semi-structured — data comes from forms, emails with predictable templates, or PDFs with consistent layouts | 2–4 months |
| 2 | Legacy system with no API; requires screen scraping, RPA bridge, or batch-file workarounds | Frequently inconsistent — requires fuzzy matching, exception handling for 20%+ of cases | 4–6 months |
| 1 | No digital interface; manual paper/fax/phone input; or requires mainframe access that IT won't open | Largely unstructured — free-text emails, scanned handwriting, or no two instances look alike | >6 months or fundamentally uncertain |
Inverted scale — higher score means lower risk. Evaluates regulatory exposure, customer/external impact, irreversibility, and data sensitivity. This is the dimension that gates AI-suitable vs. deterministic-only candidates (see the AI Risk Gate in Step 3).
| Score | Regulatory Exposure | External Impact | Reversibility | Data Sensitivity |
|---|---|---|---|---|
| 5 | No regulatory regime applies | Fully internal; error cannot reach a customer, vendor, or regulator | Fully reversible — automation can be rolled back instantly with no downstream effect | No PII, PHI, PCI, or proprietary data involved |
| 4 | Internal policy controls only (IT security, data governance — no external regulation) | Internal-facing but touches data that a downstream team uses; error causes internal rework but no external impact | Reversible within hours/days — output can be reviewed before downstream consumption | Internal proprietary data but no personal or regulated data |
| 3 | One regulatory regime applies (e.g., GDPR data residency, SOX-adjacent controls) but process is not directly in scope of the core control | Could indirectly affect customer experience (e.g., internal delay cascading to SLAs) but error is caught before direct exposure | Reversible with moderate effort — requires data correction but no permanent record issues | Contains PII or financial data but not at scale and not publicly exposed |
| 2 | Explicitly in-scope for one or more regulations (SOX key control, HIPAA, PCI-DSS) — and the automation would become part of the control | Customer or vendor-facing; an error reaches an external party directly | Partially irreversible — some actions post to systems of record and require formal correction processes | Sensitive PII, PHI, or PCI data; breach would trigger notification obligations |
| 1 | Multiple overlapping regulatory regimes with conflicting requirements; or regulator would need to approve the automation itself | Public-facing or regulator-facing; error causes reputational harm or enforcement action | Irreversible in practice — automation commits transactions, payments, or regulatory filings | Highly sensitive data at scale; breach would be material to the business |
Score the candidate process on each dimension. The weighted composite determines the pipeline tier. Hover over a score to see the anchor in the rubric above.
Composite = (Fin × 0.35 + Strat × 0.20 + Ease × 0.25 + Risk × 0.20) ÷ 5 × 100. Tier thresholds: ≥70 Priority 1, 50–69 Priority 2, <50 Priority 3 / Hold.