Every project is green until it isn't.
Twenty-two questions to someone close to the project, and you get a health score out of 100 that is calculated by rule rather than reported by the person being asked. If their own rating is sunnier than their answers, the report says so.
Fifteen minutes to answer. A number you can defend in a board paper. Free to run.
Real screens · sample project
The problem
The status report is written by the person being judged.
Not dishonestly. Optimism is how delivery gets done, and nobody wants to be the one who calls it amber first. But it means the people accountable for a project are often the last to hear it is in trouble.
Green until the month it isn't
A project rated on track for six consecutive reports does not usually deteriorate overnight. It was amber for a while and nobody had a way to say so.
Every project rated differently
One lead's amber is another's green. Without a fixed question set there is nothing to compare, so the portfolio view is a collection of opinions.
No working shown
A RAG status is a conclusion with the reasoning left out. When a board asks why, the answer is usually a conversation rather than a document.
How it runs
Three steps, and none of them is a workshop.
Someone close to the project answers 22 questions
Two sections: how the project has gone so far, and whether it is set up to succeed. Fifteen minutes, no preparation, no spreadsheet to fill in first. Most of it is multiple choice.
The score is calculated by rule, not by opinion
Each answer carries a fixed weight. The same answers always produce the same number, and unanswered optional questions neither help nor hurt. Nothing about the score is generated.
You get a summary you can put in front of a board
A score out of 100 with a band, the five dimensions underneath it, the risks the answers flagged, and the full read-back of what was actually said.
What you get
A score, its working, and the disagreement it exposes.
These are the actual screens, filled with a sample project rather than a customer's.
The score, and the five things it is made of
One number out of 100 with a band, and underneath it the five dimensions that produced it. A project can score well overall and still have a dimension in the thirties, which is usually the thing worth talking about.
- Planning & scope — How well defined the work was, and how much the scope has moved since it began.
- Delivery & budget — How the deliverables are actually tracking, and how the spend compares to the plan.
- Leadership & governance — Whether there is real executive sponsorship, and who is making the calls.
- Dependencies — The internal teams and third parties this project cannot finish without.
- Go-live readiness — Rollout, change management, data migration, and who supports it the day after.
When the rating and the answers disagree
The person filling it in still gives their own red-amber-green rating, and it is reported. It is deliberately left out of the score, because a project should be judged on what is happening to it rather than on how its owner feels about it.
Where the two disagree sharply — a project called on track whose answers land at 47 — the summary says so, and shows both. That is not a gotcha. It is the most useful conversation the assessment starts.
Why the number holds up
Calculated, not generated.
The score is the part you will be asked to defend, so nothing about it is left to a language model.
The same answers always score the same
Fixed weights, applied by rule. Run it twice with identical answers and you get an identical number. There is no run-to-run drift to explain away.
Only fair questions count
Questions with no better or worse answer — the management method, the priority ranking, who the stakeholders are — inform the report but never move the score.
The AI part is labelled
There is a written narrative, and every line of it carries an AI-generated tag. It reads the answers; it does not touch the calculation.
Who it is for
Anyone who has to answer for a project they do not run day to day.
CTOs and delivery leads
A consistent read across every project you own, so the one in trouble is obvious without you sitting through fifteen status meetings to find it.
Executives and CEOs
An independent number next to the status report you were given, and a clear flag when the two disagree.
Boards and investors
Something defensible to ask about. The score comes with its own working, so nobody has to take the rating on trust.
What it costs
Assess your first project for nothing
Create an account and assess a project. You get the score, the five dimensions, the risk flags and the full read-back, and you can share the summary with whoever needs to see it.
- No card, no sales call, no trial clock
- The whole summary, not a teaser of it
- Re-run it at a milestone to see what moved
- Paid plans add more projects and the rest of the platform
Would you rather someone ran it across a portfolio with you and talked you through what it found?
Questions
The things people ask first
Pick the project you are least sure about.
If the score agrees with the status report, you have lost fifteen minutes. If it does not, you have found something.