Signature tool · COO AI Operating Engine
Process AI-readiness scanner
List the processes you run with a few numbers you already know: monthly volume, cycle time, error rate, how rule-based the work is, how good the data is and the cost per case. The tool scores each process on value at stake, feasibility and overall AI potential, places it on a value × feasibility map, and suggests a redesign pattern — from 'automate end-to-end with an exception queue' to 'AI copilot with a human decision'. The math runs in your browser with weights you can see; AI only explains the result.
IllustrativeThe example processes and every weight and threshold are illustrative assumptions — replace them with your own numbers. Value scores are relative: 100 is the largest process in your list.
1. Your processes
Example data — edit any field or add your own processes.
2. Scores and ranking
Total annual spend $6,816,000 · 2 processes to do now ($3,096,000 a year). Value is relative: 100 = your largest process.
| # | Process | Annual spend | Value | Feasibility | AI potential | Quadrant |
|---|---|---|---|---|---|---|
| 1 | Supplier invoice matching | $1,296,000 | 74.7 | 92.5 | 82.5 | Do now |
| 2 | Maintenance work-order planning | $1,512,000 | 90.2 | 41.7 | 73.2 | Fix data or redesign first |
| 3 | Customer service: order-status contacts | $1,800,000 | 100 | 63.3 | 64.4 | Do now |
| 4 | Supplier onboarding | $432,000 | 27.6 | 54.2 | 58.6 | Cheap automation |
| 5 | Production scheduling changes | $1,056,000 | 57.6 | 48.3 | 43.2 | Fix data or redesign first |
| 6 | Quality inspection (visual, final assembly) | $720,000 | 38.5 | 73.3 | 42.2 | Cheap automation |
- #1 Supplier invoice matching: value 74.7, feasibility 92.5, Do now
- #2 Maintenance work-order planning: value 90.2, feasibility 41.7, Fix data or redesign first
- #3 Customer service: order-status contacts: value 100, feasibility 63.3, Do now
- #4 Supplier onboarding: value 27.6, feasibility 54.2, Cheap automation
- #5 Production scheduling changes: value 57.6, feasibility 48.3, Fix data or redesign first
- #6 Quality inspection (visual, final assembly): value 38.5, feasibility 73.3, Cheap automation
- Do now — High value and feasible: redesign and pilot this quarter.
- Fix data or redesign first — Worth a lot but hard today: fix the data, standardize or split the process, then pilot.
- Cheap automation — Easy but small: automate cheaply with existing tools; do not over-invest.
- Park for now — Low value and hard: leave it until data or priorities change.
3. Redesign ideas
Suggested from the rules-vs-judgment share and data availability. They are starting points for a redesign workshop, not answers.
#1Supplier invoice matching
Automate end-to-end with an exception queue
Rules-heavy work with good data: let AI and automation run the whole flow, route the cases it cannot handle to a skilled person with the full history, and retire the old manual steps.
- High error rate: add an AI first-pass check (vision, document or data validation) before the work moves on.
- Long cycle time: look for hand-offs and waiting time to remove — redesign the flow, not just the task.
#2Maintenance work-order planning
Standardize the process, then pilot AI assist
Mixed work with weak data: first reduce variants and capture the data, then add AI assistance to the most repetitive steps.
- High error rate: add an AI first-pass check (vision, document or data validation) before the work moves on.
- Long cycle time: look for hand-offs and waiting time to remove — redesign the flow, not just the task.
#3Customer service: order-status contacts
Agent drafts, human approves
A mix of rules and judgment with usable data: an AI agent prepares the case or the action, a person approves or corrects it; raise the agent's autonomy as accuracy is proven.
#4Supplier onboarding
Fix the data first, then automate
Rules-heavy but the data is not ready: digitize inputs, connect systems or add sensors before automating — otherwise the AI amplifies bad data.
- High error rate: add an AI first-pass check (vision, document or data validation) before the work moves on.
- Long cycle time: look for hand-offs and waiting time to remove — redesign the flow, not just the task.
#5Production scheduling changes
AI copilot + human decision
Judgment-heavy work: AI gathers information, flags risks and proposes options; the person decides and stays accountable.
#6Quality inspection (visual, final assembly)
Automate end-to-end with an exception queue
Rules-heavy work with good data: let AI and automation run the whole flow, route the cases it cannot handle to a skilled person with the full history, and retire the old manual steps.
Sign in to save your work privately and come back to it. Export works without an account.