Free Tool — AI Pilot Triage
Ship it, fix it, or kill it?
One AI pilot, fifteen questions, one honest verdict. The scoring isn't a sum — the verdict is your weakest dimension, and a few answers are verdicts all by themselves. A pilot can't charm its way past the thing it's failing. Runs in your browser; nothing you type leaves it.
15 questions. ~7 minutes. No email, ever. Finished verdicts are stored anonymously — answers only, never the pilot name — to publish the patterns.
Question 01 / 15
Question title
Question text
Score the evidence, not the demo.
The verdict on
Kill
Where the verdict comes from
Take the verdict to the meeting
One page, meeting-ready: the verdict, the dimension grades, the tripped wires, and the named blocker. Print it or save it as a PDF — this is the page per pilot a kill review runs on.
Generated in your browser. The pilot name never leaves it — not even for this.
One pilot is a verdict. A portfolio is a pattern — score your whole program's readiness in 10 minutes
Put the verdict where your team plans
Copy a one-line snippet with the verdict — Markdown for wikis like Confluence or Notion, plain HTML for a site or intranet page. A link in a project doc is how the next team finds this.
The snippet carries only the verdict and a link. The pilot name and your answers stay in your browser.
About this tool
What the triage does
Every enterprise has a pilot that is eighteen months old, still “promising,” and still not in production. This tool runs one AI pilot through fifteen questions across five dimensions and returns one of three verdicts: ship it, fix it, or kill it.
The five dimensions
- Value — is there a measured baseline and a number the pilot moves?
- Ownership — does a named person own the outcome, or a committee?
- Path to production — does a route into the real stack exist, with dates, or only a demo environment?
- Unit economics — does the cost per task survive contact with volume?
- Reliability & risk — error handling, failure modes, and what happens when the model is wrong.
Why the verdict is the weakest dimension
The verdict is deliberately not an average. A pilot that scores brilliantly on value and has no path to production is not a good pilot having a bad week — it is a demo. Averaging would hide exactly the dimension that kills projects, so the weakest dimension is the verdict, and a few answers act as tripwires that force it. That is also how I would triage it in a room.
FAQ
Do I need to give an email address?
No. The verdict and the dimension breakdown are shown immediately.
What if I run several pilots through it?
Do. It is built for one pilot at a time precisely so a portfolio review becomes a series of honest, comparable verdicts instead of one blended impression.
Where do my answers go?
Stored anonymously — answers only — so I can publish aggregate patterns. Progress stays in your browser.
“Kill” feels harsh. Is it?
Killing a pilot early and on purpose is a win: it frees budget and credibility for the pilots that can ship. I wrote a whole essay arguing this — it is linked from the verdict.