Process Automation Opportunity Scanner

What it does

Takes your process inventory and scores each process on automation suitability and ROI, prioritizes the backlog, and gives an architecture decision with a rationale per candidate. The core is the 5-perspective RPA suitability assessment (Syed et al. canon): each process is scored 0-5 on task · time/volume · data · system · human, weighted (0.25/0.25/0.20/0.20/0.10), and scaled to 100. An amenability veto fires: if Task < 3 or Data < 2 — i.e. the process is not rule-based or its data is not digitally accessible — the score ceiling drops to 60, so a free-text/judgment-heavy task can never score its way into "automate now."

Alongside suitability runs the ROI / Payback / NPV financial model: annual manual hours → fully-loaded labor value → Pareto happy-path automation rate (≤0.85) → error/rework savings → annual net benefit after TCO (license+maintenance), payback months, and multi-year NPV. Ranking is decided not by suitability+ROI alone but by RICE (Reach × Impact × Confidence ÷ Effort — Intercom canon). Finally each candidate is mapped via the Architecture Decision Matrix (Skill / Hook / Subagent / n8n / RPA / API / Council) to a primitive with a one-line justification.

When to use it

• "I want to automate these processes — which one first?"
• "Is this manual task worth automating, what's the ROI, how many months to pay back?"
• "Prioritize my automation backlog."
• "Should I build this in n8n, a Claude skill, or RPA?"

For the AI developer / automation / digital-transformation practitioner who wants to answer "what to automate, in what order, for what return" with numbers, not intuition.

Method / frameworks

• 5-perspective RPA suitability assessment (Syed/Geyer-Klingeberg/IRPA): five lenses, not one score.
• Automation amenability criteria: rule-based · high-volume · structured/digital data · input stability · low cognitive judgment → veto threshold.
• ROI / Payback / NPV-IRR: fully-loaded FTE × hours saved − TCO; payback typically 3-9 months for well-scoped high-volume processes.
• RICE prioritization (Intercom): a combined score sets backlog order.
• Pareto / happy-path vs exception (BPM process mining): automate the happy path, leave the exception queue to humans.
• HITL guardrail: ethical/judgment decisions (compensation, hiring, credit) are never pushed to full automation; automation drafts, a human decides.