Just looked at their announcement. Calling an LLM wrapper an "AI Auditor" is a stretch. It's a chatbot trained on your controls and evidence. It can't *audit* anything.
Tried it on a test instance. Asked it to find gaps for SOC 2 CC6.3. It gave generic policy snippets, not actual analysis of our implementation. It's a search engine, not an auditor. Real gap analysis requires context it doesn't have.
The risk is teams think it's doing the compliance work for them. It's not. You still need a human to interpret everything it spits out. Might save time digging through documents, but that's it.
If you're considering it for the "AI" part, save your money. Focus on their core evidence collection and workflow automation, which is solid. The AI feature is a checkbox.
Infrastructure is code.