Your members are talking to your AI. Do you know what it's telling them?
Cimbiotic is an independent third party. We test the chatbots and agents your members rely on, from the outside. We keep every reply as evidence and give your Board an Assurance Statement it can check.
Picture this email. It could arrive on any Tuesday.
“Your assistant told me the registration deadline was the 15th. Your website says the 22nd. Which is right?”
Hypothetical, but not far-fetched.
Your first question is not which date is right. It is which assistant said that: the one on your website, the one in your event platform, or the one in the help desk? Which version? And what changed since anyone last checked?
Buying AI is the easy day. Every day after, vendors update their models, staff edit instructions, and nobody tells you. Most associations do not have one vendor. A membership system, an event platform, a help desk and a website chat can each have their own assistant, each changing on its own schedule.
A launch test describes launch day. A vendor’s own results come from the party that sells the product. Neither shows that the assistant your members meet today is still on the guardrails you set.
Not sure what AI you are running? Take the free Scorecard.
A question lands on your desk.
- A Board member asks whether anyone has checked the member assistant.
- A vendor renews, and its assurances describe the platform, not your policies.
- A member reports a wrong answer, and you need to know whether it was a one-off.
- An insurer, auditor or RFP asks how you govern AI.
If none of these sound like you yet, the Scorecard is a calm place to start. Take the free Scorecard.
If one of these has landed on your desk, a 30-minute call is the quickest next step.
Book a 30-minute callAn independent check, on a regular schedule.
CimAssure tests your chatbots and agents from the outside, the way a member would use them. We keep every reply as evidence and give your Board an Assurance Statement it can check. The tests repeat on a regular schedule, and again when the AI changes.
See how CimAssure worksSee a sample Assurance Statement
The sample is illustrative: a fictional association and assistant, with invented findings.
Our findings are based on the evidence.
No vendor pays us, and no fee depends on what we find. We take no commissions, referral fees, equity or board seats from any vendor. Our criteria are published on the Framework page. We reserve the right to fail an assessment and withdraw a mark.
Start from your situation.
Your Board has asked, and you need an answer soon.
AI Red Teaming. An independent, outside-in test of your chatbot or agent, run once. We probe for prompt injection, jailbreaks, data leakage and unsafe outputs, keep every reply as evidence, and rate the result for your Board. It needs no Baseline, and the fee is credited toward your first year of CimAssure.
See AI Red Teaming →Your assistant is live, and it needs to stay trustworthy.
CimAssure. A three-week Baseline establishes what AI is running, who owns it and what it must not do. We then test on a regular schedule, and again when the AI changes, so your Assurance Statement stays current. Agent testing is available now.
See how CimAssure works →You are not sure what AI you are running.
Free Trust Readiness Scorecard. Seventeen questions, about five minutes, and you see your results without signing up.
Take the free Scorecard →