Score one workflow across process, data, technology, and ownership. Use the result to decide whether to build, pilot, or fix the foundations first.
No signup. Your answers never leave this page.
Answer for one workflow, not the whole company. The result is a heuristic, not a guarantee.
Every number this tool shows, and where it comes from.
Scoring. Each question is worth 0, 5, or 10 points (No / Sometimes / Yes). Total score 0–100. The four categories (Process, Data, Tech, Org) carry different max weights based on how much each tends to predict deployment success.
Verdict bands. 85+ ship-ready, 70–84 almost there, 50–69 pilot first, 25–49 foundations needed, 0–24 not yet. The bands are calibrated against patterns we’ve seen across Orbit deployments and AI failure post-mortems published by McKinsey, BCG, and the MIT Sloan AI study.
Why these 10 questions. 70% of failed enterprise AI projects fail at process, data, or organizational fit, not at the model. The questions front-load those gates. References: MIT Sloan State of AI in Business, McKinsey State of AI.
What it doesn’t do. This is a heuristic, not a guarantee. It can’t see your codebase, your customer contracts, or your team dynamics. Use it as a conversation starter with the people who will own the system, not as a final answer.
Question set last reviewed 2026-05-06.
The rest of the toolkit runs in your browser too.
Keep the first version on Approve-each and use the run log to decide what earns autonomy.
500 trial credits. Card required. $0 charged today.