Skip to content

The Record · Part 4 of 5

What Underwriters and Sureties Now Ask About AI

Few firms can answer with documents. That’s the opening.

By Jared Smith, Founder & CEO, SymphonIQ ·  · 4 min read

The Record, Part 4 of 5: What Underwriters and Sureties Now Ask About AI. Few firms can answer with documents. That’s the opening.

Sixty days before renewal, your broker forwards a supplemental questionnaire: extra questions on top of the usual application. This one is about AI.

Which AI tools do you use? Who approved them? How are staff trained?

The honest answer is “we’re not sure.” (This scenario is illustrative.)

This issue is about giving a better one.

The questions are specific

In its July 2026 market update, the broker Lockton says some carriers now ask “increasingly granular questions about how AI is used.” They want to know which tools are approved, who may use them and for what.

They also ask whether the use is agentic, generative or both. Generative AI creates content when someone asks for it. Agentic AI can also take actions on its own.

Lockton lists questions like these:

  • “What policies exist that govern AI’s use?”
  • “How are employees trained on AI?”
  • “How are data and privacy protected?”
  • “How is AI addressed in contracts with vendors?”
  • “Are there human resources use cases for AI?”
  • “Is there a way to escalate complaints about AI use within the company?”

Lockton’s advice: start the renewal conversation with a clear account of how you use AI and the guardrails around it.

Underwriters are short on data

Underwriters are the people who decide whether to cover you, and on what terms.

The Lloyd’s Market Association represents the underwriting businesses at Lloyd’s of London. It surveyed 144 market participants about AI and found the responses were “informed by views and opinions rather than data in most cases.”

When the underwriter is working from opinion, your documents become the data.

Few firms can show their work

Few firms have those documents ready. That’s the opening.

In IBM’s 2026 breach study, only 32% of breached organizations had an AI governance policy, meaning written rules for how AI is used.

RICS is the global professional body for surveyors. In its 2026 survey, only 14% of construction respondents were strongly or fully aligned with its AI standard.

A firm that answers with documents stands out.

Your answers become evidence

Standing out only helps if every answer is true.

The statements on your application are called representations. The insurer relies on them when it writes your policy. Lawyers at Pillsbury who represent policyholders warn that representations about AI “capabilities, governance and controls should be scrutinized carefully.”

That scrutiny matters twice: on the application, and in any coverage dispute later.

Answer accurately. Don’t describe controls you don’t have.

Sureties are watching too

It’s not just insurers. Carl Oliveri of Grassi argues that technology investments “should be considered in the surety’s underwriting strategy.” He wrote that in Surety Bond Quarterly, published by NASBP, the trade group for surety bond agents and brokers.

Merchants Bonding, a surety company, gives contractors a simpler rule: “Technology should assist decisions, not replace them.”

If AI touches hiring

One of Lockton’s questions asks about HR use. AI in HR comes with its own laws.

Colorado replaced its original AI Act with SB 26-189. The new law takes effect January 1, 2027, with narrower notice and disclosure duties.

If any AI tool screens or ranks candidates, get counsel involved.

The AI underwriting binder

Put your answers in one binder. Here’s each question and the document that answers it:

  • Which tools are approved, and who may use them? An approved-tool list with account types (business or personal) and permitted uses, including tools at outside firms.
  • Agentic, generative or both? A tool inventory that notes what each tool can do on its own, such as send, post or change records.
  • What policies govern AI use? A written AI use policy with a named executive owner and an approval record for each tool.
  • How are employees trained? Training records and signed acknowledgments.
  • How are data and privacy protected? A restricted-data rule (what never goes into AI), training turned off in each tool’s settings, and proof you monitor and audit for unapproved AI.
  • How is AI handled in vendor contracts? Signed AI riders (contract add-ons that cover AI use) and a schedule listing each firm’s approved tools.
  • Any HR use cases? A list of every HR use of AI, with counsel review of any hiring tool.
  • How are complaints escalated? An incident procedure, an incident log and a record of corrective actions.

What it means for a contractor

You can’t control the questionnaire. You can control whether every answer has a document behind it.

Do this month

  • Start the binder now. Have it ready 60 days before renewal, and walk your broker through it.
  • Have counsel or your broker review every AI answer on the application before you sign.

Ask your surety

Your surety’s prequalification review is how it decides whether to bond you, and for how much. Ask:

“Is AI governance part of your prequalification review, and what would you want to see?”

Next week

Part 5, the finale: subs, consultants and the contract chain, plus a 90-day plan.

Part 5 publishes Tuesday, November 17, 2026, at 7:45 a.m. Mountain. Get it by email · Subscribe on LinkedIn · RSS feed

Free resource: Our AI risk framework for contractors maps each underwriter and surety question to the document that answers it.

Get The Record by email

Each part in your inbox the morning it publishes.

Then the occasional new guide or template for contractors. Unsubscribe in one click.

We confirm your address by email first. Privacy Policy

General information, not legal or insurance advice. Talk to your broker and counsel.

Sources

  1. Lockton Spotlight: AI, July 2026
  2. CSIS: The Insurance Industry’s Retreat from AI (Lloyd’s Market Association survey), Sept 4, 2026
  3. Pillsbury: AI Exclusions in Insurance Policies, Apr 13, 2026
  4. Surety Bond Quarterly (NASBP): Oliveri, How Contractors Leverage AI, July 2025
  5. Merchants Bonding: 4 Safeguards for Contractors Working With AI
  6. IBM Cost of a Data Breach Report 2026 (PDF), July 29, 2026
  7. RICS AI in Commercial Property and Construction Report 2026, Aug 18, 2026
  8. Ropes & Gray: Colorado scales back AI law, May 19, 2026

About The Record

The Record is written by Jared Smith, founder and CEO of SymphonIQ, which builds governed, single-tenant AI for engineering and construction.

Subscribe on LinkedIn Book a working session
Book a working session