Free · No email needed to run it · About ninety seconds
Agentic Readiness Diagnostic
Assess one workflow. Get three readings, a named constraint, and an honest answer, including the answer that you should not do this.
Three readings rather than one score.
A score is less useful than the truth, because one number is being asked to answer three different questions at once.
Opportunity
Is there enough value here to be worth pursuing?
Deployment readiness
Can this organization actually ship it?
Safe autonomy
How independently should an agent be allowed to operate?
On top of those, the assessment returns a decision, the single constraint that is binding you, a recommended autonomy mode, an estimated range of economic headroom with a stated confidence, and a computed answer to what would change the result.
Example output
One workflow assessed
- Opportunityvery high
- Deployment readinesslow
- Safe autonomyhigh
- Decisiondo not build yet. Establish ownership and a baseline first.
- Binding constraintno accountable workflow owner.
- What would change thisassigning one owner with authority to change the workflow moves deployment readiness from low to moderate. Establishing a measured baseline would then produce a decision of proceed with constraints.
Two things it does differently
- “I don’t know” is a first-class answer.You are never made to invent certainty. Unknown is not zero. It lowers the confidence of the assessment, and sometimes it is the finding: we cannot tell whether this workflow has economic headroom, because nobody currently knows what it costs, and that is the first thing to fix.
- It knows where its competence ends.Where legal, regulatory, contractual, security or data-residency restrictions materially limit automated action, the output is “requires specialist review”, not a recommendation. A web questionnaire has no business adjudicating financial or medical regulation.
The label
Carried on the tool and on every result
The diagnostic encodes a practitioner heuristic, not a validated model.
Ownership and measurement are expected to be disproportionately important gates in the current heuristic. Testing whether that assumption is correct is one purpose of our research and of the outcome tracking behind this instrument.
Every assessment is versioned. A result produced under one version of the instrument is never silently rewritten under another.
Before it opens: independent certification
The instrument is built and has been audited against sixteen adversarial cases designed to break its reasoning, with the expected output written down before each one was run. It goes public after independent certification, not before.
Opens your email client, addressed to a person: hello@900labs.com