AIUC-1 certification: how an AI agent audit links to insurance

What is AIUC-1?

AIUC-1 is a standard for AI agents that covers security, safety and reliability. It is published by AIUC, the Artificial Intelligence Underwriting Company, which audits companies against it and ties insurance terms to the outcome¹.

AIUC describes itself as a frontier AI standards and insurance company, based in San Francisco, that certifies and insures AI agents. Its homepage says Cursor, ElevenLabs and Harvey helped build AIUC-1¹.

The idea is to give insurers, and the agent's customers, independent evidence of how an agent behaves before anyone prices the risk. AIUC's site lists companies as certified without saying they are insured, so certification alone does not show that a policy is in place¹.

Who is behind AIUC

AIUC emerged from stealth in July 2025 with a $15M seed round, according to Insurance Journal². In our register it is a standard setter, not an insurer.

Sources conflict on its later funding. AIUC's homepage describes a $55M Series A¹, while Coverager reported on 15 September 2026 a $40M Series A led by Ribbit Capital on top of the $15M seed, which also comes to $55M³.

How does AIUC-1 link an audit to insurance?

The audit comes first and the insurance terms follow from it. AIUC says it links insurance terms to the audit outcome, so the terms on offer depend on how the agent does against the standard¹.

• A company that builds or deploys AI agents is audited against AIUC-1¹.
• The audit result shapes the insurance terms offered for the agent¹.

AIUC's ElevenLabs update says unnamed insurers offer an AI-specific policy that can underwrite the actions of agents deployed by ElevenLabs' customers, for risks such as an agent giving a customer incorrect information⁴. Other descriptions add exposure of customer personal data and misquoted advice^{1, 5}.

What is not public: the insurer behind AIUC-1-backed policies, the wording, the limits and the prices. Insurance partners were unnamed in 2025 coverage², and AIUC's homepage and its ElevenLabs update refer to insurers without naming any^{1, 4}.

How does a company get AIUC-1 certified?

Through AIUC. Certification means independent testing of the agent against the AIUC-1 standard, and AIUC links the insurance terms to that result¹.

Our register records no published fees, timelines or renewal terms for the audit, so ask AIUC directly. ElevenLabs, the first company to announce an AIUC-1-backed policy, says its own customers can start the process through a form on its site, and that its agents went through 5,835 technical tests across 14 risk categories⁶.

Expect the audit to cover the same ground underwriters ask about for any agent: what it can do without human sign-off, which systems it can reach, how its outputs are tested and logged, and how quickly it can be stopped. That is our general expectation, not AIUC's published method.

Which companies have AIUC-1 certification or cover?

One company has said publicly that it carries an AIUC-1-backed policy. AIUC's site lists others as certified without saying they are insured.

• ElevenLabs announced on 12 February 2026 that it was the first to go live with an AIUC-1-backed policy, for its voice agents. No insurer, limits or terms were disclosed^{4, 6, 7}.
• Cursor is listed on AIUC's site as AIUC-1 certified, and AIUC calls KPMG LLP the first Big Four firm with an AIUC-1 certified AI capability, aIQ Capture. The pages we read do not say either is insured¹.

• AAIUCAIUC-1-backed AI agent insuranceLimit not publishedThrough AIUC-1 certification
• AAIUC and BeazleySpecialist AI liability product on Beazley paper (trade press)Limit not publishedHow to buy not stated

The first listing is confirmed on AIUC's own pages. The second rests on trade press alone, as the next section explains.

Is AIUC working with Beazley?

It has been reported, not confirmed. Trade press reported in May 2026 that AIUC secured underwriting capacity from Beazley to build a specialist liability product for AI companies⁸.

The only write-up we could read is an aggregator piece that does not cite its original, and it gives no limits, pricing, customers or launch date⁸. Neither AIUC's homepage nor its ElevenLabs update names Beazley or any other insurer^{1, 4}. We list it as trade press only until AIUC or Beazley confirm it.

Separately, Beazley sells three AI endorsements on its own Full Spectrum Cyber policy, confirmed on its own pages⁹. The AI liability insurance guide covers those.

How does AIUC-1-backed cover compare with other AI agent insurance?

AIUC-1 is one of several ways insurers get evidence about an AI system before covering it. These are the others in our register:

• Armilla: each policy comes with independent AI system certification and risk reporting, with limits up to $25M per organization¹⁰.
• Ollive: pairs runtime monitoring of AI agents with planned liability cover for AI vendors. Its page says the insurance program is being designed¹¹.
• Mayflower Specialty: scores AI-specific risks such as bias, drift and hallucination for an affirmative AI liability program issued by Hadron Specialty Insurance Company^{12, 13}.
• Google Cloud: pairs a free security scan with cyber policies from partner insurers, and announced affirmative AI cover as available in May 2025 without naming the insurer that writes it¹⁴.

The difference is that AIUC-1 is a published standard aimed at AI agents, while the others assess systems as part of their own underwriting. The AI agent insurance guide compares the agent covers, and the AI startup insurance guide shows where this fits for a young company.

What should you ask an AI agent vendor about AIUC-1?

Ask two separate questions: is the agent certified, and is it insured? AIUC's site lists certified companies without saying they are insured, so one does not prove the other¹.

• Ask for a certificate of insurance that names the insurer, the limit and the policy period.
• Ask whether the policy names AI or agent errors, or is silent on them.
• Check your contract with the vendor: who pays if the agent harms one of your customers?
• Check your own cover too. Your E&O or cyber policy may respond first, or may exclude AI.

For your own policy, the coverage checker shows where to look. The methodology page explains how we decide whether a listing is confirmed, reported or announced.

AI Insurance Tracker does not sell insurance or certifications. If a customer asks you for AIUC-1 or for AI cover, a licensed broker can tell you what is available for your agent.