AI "employee" platform Aivah discloses no insurance or compliance certification

Aivah, which sells configurable "AI employees" that answer customer questions and run sales and support tasks, publishes no compliance certification, regulatory alignment or insurance disclosure anywhere on its site, according to an AI Dispatch review completed Oct. 3. The access and oversight controls it does offer are set before an AI employee starts working, not enforced while it runs.

The company lets a business "choose the tools and permissions your employee needs for the job" and "choose the sources, review tool permissions and test the work" before deployment. It also lets a business "define when a customer success manager takes over" from the AI employee. Each of these controls is configured in advance; the site describes no limit on scope, spending or authority that is checked once an AI employee is handling a conversation.

On record-keeping, Aivah's dashboard lets staff "filter by employee and date to review usage and recent activity" and "open the relevant chats, leads, quizzes or calls to understand what happened." That gives customers a reviewable record of what an AI employee did, but the site does not state whether the record is tamper-evident, how long it is kept, or whether it can be exported.

The company's site contains no mention of SOC 2, ISO 27001, GDPR, HIPAA or any other compliance standard, and no mention of insurance, indemnification or liability coverage for the actions its AI employees take.

Aivah's entry-level Hobby plan is $60 a month, or $576 billed yearly, for one AI employee, one "Agentic Presenter" avatar and 50,000 shared monthly credits — a flat subscription with no pricing tied to a verified result. Customers can choose among four foundation-model families: OpenAI, Anthropic, Google and xAI.

AI Dispatch scored Aivah under its published trust-gap rubric, which rates vendors zero to three on five dimensions with evidence required for every score. Aivah scored 0 of 3 on compliance, outcome settlement and insurance/indemnity — three of the five dimensions with no public disclosure at all — and 1 of 3 on both audit trail and runtime governance, reflecting controls that exist but are set in advance rather than enforced while the system runs. The rubric is published at aidispatch.news/methodology.