Try VALUE-AI on a real idea. Then tell us what you think.
VALUE-AI is a free framework for deciding whether an AI idea is worth building, and what has to be true before it scales. Pick a use case from your own work, take the assessment, read the report, and give us five minutes of honest feedback.
5–15 minutesNo sign-inYour answers stay anonymous unless you leave an email
- 1Think of a use caseA process you'd like to improve, or an AI idea you've been asked to evaluate.
- 2Answer the questionsBusiness leads: about 20 in plain language. IT and AI leads: a quick scan or the full assessment.
- 3Read your reportVerdict, scores, the weakest link and what to fix next.
- 4Tell us what you thinkA short survey on the framework, the process and the report.
What VALUE-AI is
Five questions every AI idea has to answer
Each letter is a pillar. Together they cover sixteen areas, from whether the problem is real to whether a pilot can prove anything. You answer facts about your situation; the framework does the scoring.
Is there a real, owned, measurable business problem worth solving?
Is AI genuinely the right solution — and is it feasible with the data and systems we have?
Can we run it safely, lawfully, and accountably — before, during and after deployment?
Are the people, workflows and change plan ready for it to actually be used?
Can we prove value with a controlled Proof of Value before we scale?
Decide before you build
Factual questions — never self-ratings — tell you whether an idea is worth a Proof of Value, needs work first, or isn't an AI problem at all.
Risk can veto value
A high score can't hide a serious risk. Weak controls on a high-risk use case cap the recommendation, whatever the total.
A report you can take to a decision
Verdict, scores across five pillars and sixteen dimensions, the weakest link, and the questions to resolve next.
New in v1.4
- Questions for AI agents that act on their own — tool access, actions that can't be undone, hidden instructions, hard limits, traceability.
- Ready for Australia's automated-decision disclosure rules, which start on 10 December 2026.
- Vendor terms, shadow AI, and agreeing your stop criteria before a pilot starts.
- Now draws on Australia's Guidance for AI Adoption, Singapore's agentic-AI framework and OWASP's AI threat lists, alongside NIST and OECD.
What to expect
The whole thing, in 90 seconds
A use case goes in, questions picked for that use case come out, and the report follows straight away. A narrated run through a quick assessment, step by step. Captions are on if you’re watching without sound.
Want to go deeper first? The framework page explains every pillar, the oversight levels and the scoring rules.
Quick reference
Take the framework with you
Free downloads, no form to fill in.
Your turn
Bring one use case to mind
Pick something real you’re weighing up — a tool you’re being sold, a pilot someone wants to run, an agent your team is excited about. Answer as things stand today, not as you hope they’ll be.
Stuck? Borrow one:
The VALUE-AI Advisor can run the same assessment as a chat, explain any question as you go, and walk you through the lifecycle templates. It’s the Advisor from the full platform.
Done this before? to see your saved assessments.