Hard questions, answered plainly
Including the one about whether this is a certification. It isn't, and we say so in the same words every time.
Those tools are run by the team that built the AI, on data they chose. Useful for the builder, but self-graded. We are independent: a separate team, our own tests, private held-out cases, and the full method shown.
Every score ships with its working: what was counted, out of how many, on which held-out set, and a breakdown of the failures. If we cannot show how a number was made, we do not publish it.
No. We provide an independent evaluation report and scorecard — an evidence-based opinion at a point in time. It is not a certificate, a guarantee, or an accreditation, and we do not present it as one.
All three. Agents are audited and tested across their whole loop — understand, observe, locate, act, verify. AI solutions are evaluated on held-out tasks and under attack. AI programmes are assured as a portfolio, including the policies and inventory that govern them.
A clear description of what the AI is meant to do, access to run it, the claims being made about it, and examples of the real tasks it should handle. The request form asks exactly those four things, so the first reply can already contain a draft scope.
It depends on the work — how many workflows are in scope, how complex the agent is, how we get access, and how deep the red-team goes. Every engagement starts with a written scope that fixes the timeline and the price before anything begins.
You get a ranked list of the specific issues and how to reproduce each one, so the builder can fix them. Then we re-evaluate.
Yes. We work under NDA as standard. Your data, prompts and outputs are used only for your evaluation and seen only by the evaluation team. Our held-out test sets are never shared with the vendor being evaluated. At the end of an engagement we return or delete your data unless you ask us to keep it for re-evaluation.
Every engagement is scoped individually, so we don't publish a price list. Tell us what the AI does and what is in scope; you receive a written scope with a fixed price and timeline. Write to hello@wisdombridge.tech with the subject line “Engagement”.
It depends on the volume of the work. A single agent or AI solution is scoped to that system; an enterprise programme runs over multiple months. The length is evaluated during scoping — from the number and complexity of the systems in scope — and fixed in the written scope.
Yes, through Remediation Support. It is delivered by a team kept separate from the evaluators, and the people who help fix a system never re-evaluate that same system. That separation is what keeps the re-evaluation independent.
WisdomBridge-qualified practitioners with backgrounds in AI governance, information security, LLM evaluation and quality engineering. We don't publish individual names on the site; the backgrounds and credentials of the people assigned to your engagement are shared under NDA during scoping.
No. The report belongs to whoever commissions it. A vendor may choose to share its report with buyers; we never publish client results or name clients without written permission.