AI Risk Insights Platform · How It Works

Four steps. One AI read.

Order AI Risk with your normal scan, and CyRisk gathers public evidence, checks it against the broker’s documents, and delivers a report built for the bind decision.

Step 1

Order it with the scan you already run.

AI Risk sits next to Security and Privacy on the scan form. Select all three for a full read. You can also upload the application and any supporting documents; they are saved with the company so later scans can reuse them.

Step 2

Gather evidence from public sources.

CyRisk looks where AI use actually shows up, and keeps only evidence it can date and cite.

The company’s own website

Product and AI pages, trust and security centers, privacy policies, terms, data processing agreements, sub-processor lists, developer documentation and careers pages.

The public record

Recent news, supplier and partner pages, company events and, for public companies, AI and cyber disclosures in regulatory filings.

Legal and regulatory matters

Lawsuits, regulator actions, intellectual property disputes and reported incidents.

CyRisk’s own scans

Recent CyRisk Security and Privacy results for the same company.

The broker’s documents

The application and supporting documents, when uploaded.

Sources it leaves out

Social media and job boards are deliberately excluded.

Step 3

Analyze use, depth and risk.

How the company uses AI

Customer-facing assistants, AI that makes decisions about people, internal productivity tools, or building AI for others.

How deeply it is built in

Whether AI is peripheral, integrated or core, how much it acts on its own, and whether the company builds models or uses someone else’s.

Where the risk sits

Each AI risk area is rated Not detected, Low, Moderate, High or Critical and tied to the coverage line it affects.

Cross-checks. The report compares what the application says with what the public evidence shows, and flags contradictions, unsigned applications and documents that name a different company.

Step 4

Deliver where you already work.

AI tab

Results appear on the company in the CyRisk platform.

Company list badge

See AI risk at a glance across your submissions.

AI Risk Report

A full report as HTML or PDF.

Executive report

An AI card in the combined Security, Privacy and AI executive report.

Rules of evidence

Sourced, dated and honest about gaps.

  1. Every finding links to its source, the date it was read and the exact quote. Findings that can’t be traced to a source are dropped.
  2. If the company’s website can’t be read, the scan fails openly rather than returning a false clean result.
  3. “No AI found” means none was found in the sources reviewed, not that the company uses no AI.
  4. Findings drawn from your uploaded documents stay within your account, and deleting a document removes everything derived from it.

Questions underwriters ask

Frequently asked questions.

What does CyRisk need to run an AI scan?

A company domain with a readable English-language website. Uploading the application and supporting documents is optional but lets the report check them against public evidence.

Does the AI scan use social media or job boards?

No. Those sources are deliberately excluded.

Is the AI scan a continuous monitor?

No. Each AI scan is a point-in-time read of one company, run alongside Security and Privacy when you order it.

Who can see findings from our uploaded documents?

Only your account. Public findings can be shared, but anything drawn from your documents stays private, and deleting a document removes what came from it.

See AI risk on your own submissions.

Get the sample AI Risk Report, or book a walkthrough with our underwriting team.