Spectralgraph
Insurers

AI testing evidence, ready for your AIS program.

Oklahoma's Bulletin 2024-11, like the NAIC model bulletin it follows, expects insurers to govern the AI systems they use and to be able to show it. Spectralgraph is an independent AI-testing laboratory that measures fabrication risk in the model you host and delivers dated evidence your program can stand on.

Request the sample report See the public index
FIXED PRICE · FIVE BUSINESS DAYS · NO MEETINGS, EVERYTHING IN WRITING · NO CUSTOMER DATA
The regulatory moment

The bulletin asks for governance you can show, not governance you can describe.

Under OID Bulletin 2024-11 and the NAIC model bulletin pattern it adopts, an insurer using AI systems is expected to maintain a written program with controls proportionate to the risk of each use. Programs built from policy documents alone tend to share a weakness: when someone asks how the model itself actually behaves, the file has adjectives where it needs numbers.

It's not a checklist or a vendor questionnaire. What it does is measure the model itself, reading the model's own internal signals rather than grading its response output. Using your real question types on a model you host, it maps where the model fabricates, how often, and on which topics, down to specific flagged answers your team can verify with their own eyes. The deliverable is a signed, dated screening measurement built for your governance file. In testing so far it catches fabricated answers at about 0.85 AUC on models it has never seen before; that figure is provisional until the work is published.

What it needs from you

No policyholder data. Nothing installed. Five business days.

The test needs a sample of your real question types, claims, coverage, and service questions as your teams actually phrase them, and a model endpoint you designate. It never touches policyholder data. In scope are models your company hosts or controls, including private Azure OpenAI and AWS Bedrock deployments; vendor-attested SaaS AI is out of scope and we will say so.

The first measurement is your baseline; quarterly re-measurement builds the trend record that makes an AIS program inspection-ready as models change. For the provision-by-provision detail, read the full crosswalk to Bulletin 2024-11: what the report addresses, what it supports, and what stays honestly outside scope.

Terms

Fixed fees, written engagement, nothing success-based.

Fees are fixed and never success-based: the LLM Validation Report is $9,500, Continuous Assurance is $6,500 per quarter for one model, and the pre-deployment Model Selection Study runs $4,500–$7,500 by candidate count. Full details and the founding client program are on the home page. Every engagement runs under a written agreement before any work begins.

Straight answers

Questions insurance compliance teams ask first

Does this satisfy OID Bulletin 2024-11?
No single vendor artifact satisfies a governance bulletin, and we will not pretend this one does. The bulletin asks insurers to run a written AI systems program with governance and controls proportionate to risk. What the measurement provides is a concrete piece of evidence inside that program: an independent, dated record of how your model actually behaves, which is the kind of artifact that makes the rest of the program credible.
Does the test touch policyholder data?
Never. The instrument needs a sample of your real question types and a model endpoint you designate. Policyholder data stays where it is.
Our AI is inside vendor software. In scope?
If the vendor hosts and attests to the model, there is nothing of yours to measure and we will say so. In scope: models your company hosts or controls, including private Azure OpenAI or AWS Bedrock deployments and fine-tunes.
What lands in the governance file?
A signed, dated screening measurement: where the model fabricates, how often, on which topics, with the specific flagged answers your team can verify line by line. Quarterly re-measurement adds a trend record as models and prompts change.

See the deliverable before you commit to anything.

Reply and we will send the sample validation report so your compliance team can judge the artifact itself, not a brochure about it.

Email the lab
REPLIES SAME DAY · EVERYTHING ANSWERED IN WRITING, ON THE RECORD