Evidence you don’t have to take our word for
Prove it · 05 · For internal audit and risk teams
The auditor asks a simple question: how do you know this log is complete and unaltered?
The honest answer, for most AI tools, is that the vendor says so.
Fifty percent of organisations cannot produce a complete AI data access audit record within one business day. [1] And nearly all (98%) say they conduct a formal AI assurance review at least annually. [2] Put those together and a lot of annual assurance is resting on records that are slow to assemble and held by someone else. If you can’t produce the record when it’s asked for, the audit finding is yours, not the vendor’s.
Who keeps the record matters
When the AI tool that acted also holds the only record of what it did, it’s marking its own homework. The record may be accurate, but the only way to check it is to ask the people it’s meant to hold to account.
Internal audit already knows this pattern. We don’t accept a business unit’s own summary as evidence that its controls worked. We shouldn’t accept an AI tool’s own record either.
Without MeshLogic, the only way to check the log is to ask the supplier of the tool that wrote it. With MeshLogic, your auditor checks the record themselves, and where the evidence is missing, the record says so.
What should evidence you can rely on look like?
- Captured where it happened. Recorded on the device where the agent acted, at the time it acted.
- Locked. Each record is locked so any later change shows.
- Held outside the vendor. A copy is kept outside the vendor’s control, so no one party can quietly rewrite history.
- Checkable by you. Your auditor or regulator can check the record themselves, without asking the vendor.
- Honest about gaps. Missing evidence is reported as insufficient, never filled in.
That gives your auditor evidence they can rely on, without waiting on the vendor or having to trust it.
Where MeshLogic fits
MeshLogic is a vendor too. The difference is that you don’t have to take our word for the evidence we provide. MeshLogic sits apart from the AI tools it records. Each record is created on your organisation’s own Windows and Mac computers at the moment the agent acts, and sealed there, a bit like a tamper-evident evidence bag: if anyone changes it later, including us, the seal shows it. A copy is also kept outside our control. That’s why you, your auditor or your regulator can check any record yourselves, without asking us. Our evidence chain has been independently verified, and when we can’t prove something, we say so.
MeshLogic is SOC 2 Type 2 across all five trust services criteria, AWS Qualified Software and live on AWS Marketplace.
References
[1] Kiteworks, 2026 AI data governance and security survey, press release, 29 July 2026. https://www.kiteworks.com/company/press-releases/kiteworks-2026-ai-data-governance-security-survey/
[2] EY, AI Risk and Governance Survey, press release, 15 September 2026. 202 senior AI decision-makers at publicly traded companies with $1B+ annual revenue. https://www.ey.com/en_us/newsroom/2026/09/ey-survey-finds-that-autonomous-ai-implementation-outpaces-oversight-yielding-an-ai-governance-gap
Want to test your AI controls before the next board paper?
We are taking on a small number of design partners. Bring a control list and we will show you what the evidence carries and what scores insufficient. We can start under NDA.