How are insurance agencies using AI?

Vertafore surveyed more than 1,300 independent agency professionals for its 2026 Agency Trends Outlook. The sentiment split is more interesting than the headline, and one number from the client side of the same research changes what you should build.

The sentiment split

Asked about AI, agency professionals landed in four camps.

How independent agency professionals describe their stance on AI

More than 1,300 respondents. Percentages as reported.

Curiousexploring use cases
39%
Cautiouswaiting for proven results
29%
ConfidentAI will transform the work
23%
Skepticaldoubts current value
9%
Table view
StanceShare of respondents
Curious — exploring use cases39%
Cautious — waiting for proven results29%
Confident — AI will transform everyday work23%
Skeptical — doubts AI's current value9%

Source: Vertafore, 2026 Agency Trends Outlook. Survey of 1,300+ independent agency professionals.

Only 9% are actually skeptical. Add curious to confident and 62% are already leaning in. The largest single group — the 39% who are curious and exploring — plus the 29% waiting for proof, are the same audience described twice: people who will adopt when someone shows them a use case that survives contact with their own book.

Nearly two-thirds of the industry is not asking whether AI works. They're asking which part of their day it is allowed to touch.

What the same research says separates high performers

Asked what distinguishes the agencies doing well, respondents pointed first at something unglamorous.

What agency professionals name as the high-performer differentiator

Two most-cited answers.

Client communication
52%
Streamlining operations
39%
Client communication Streamlining operations
Table view
DifferentiatorShare naming it
Client communication52%
Streamlining operations39%

Source: Vertafore, 2026 Agency Trends Outlook.

Communication first, operations second — and both are precisely where an agency's headcount caps out. You cannot communicate with more people than you have hours, which is the entire reason the 39% are curious in the first place.

And then the number that should shape the build

From client-side research referenced in the same outlook: 85% of policyholders want transparency about their agent's use of AI.

85%
of policyholders want to be told when AI is involved
Client survey referenced in Vertafore's 2026 Agency Trends Outlook.

That single figure disqualifies an entire category of product. Any AI that works by impersonating a producer — using their name, answering as them, hoping nobody asks — is building on the opposite of what 85% of clients say they want. It is also, separately, the approach most likely to produce a complaint you can't defend.

The alternative isn't weaker. An AI that says plainly that it's the assistant on a named producer's team can still open the conversation, answer the price question, handle the brush-off, and book the appointment. It gives up exactly one thing — pretending — and in exchange it satisfies the 85%, survives the awkward question, and keeps a licensed human where the licence is legally required to be.

Where this leaves the cautious 29%

"Waiting for proven results" is a completely reasonable position for a business whose product is a regulated promise. What it usually means in practice is three specific worries, each of which has a concrete answer:

  • "It'll say something I can't defend." Bound what it may discuss. Coverage, underwriting, and eligibility questions go to a licensed human — not as a fallback, as a rule.
  • "It'll burn my phone numbers." Quiet hours in the lead's own timezone, STOP honored instantly and everywhere, opted-out records failing closed on every send path, and message variety so carriers don't fingerprint one repeated template.
  • "I'll lose control of the relationship." Per-conversation and global pause. New behaviour that arrives as a proposal a human approves rather than a change that happens to you.

Those are the terms on which the curious 39% become adopters — and they're the terms Virtual Closer was built to meet, because it was built inside an insurance agency where getting them wrong had consequences for the people in the room.

The gap isn't enthusiasm — it's readiness

The same body of research contains a second finding that explains why the curious 39% have not become adopters. Asked how prepared their agency was to keep pace with technology and market change, 46% said only somewhat or not at all — and at agencies of six or fewer staff that rose to 55% (Vertafore, building confidence in your agency's tech stack).

Read those two findings together and the picture is specific rather than vague. This is not an industry that needs persuading that AI works. It is an industry where most firms do not currently have the operational footing to adopt anything — and the smaller the agency, the truer that is, which is the opposite of the usual assumption that small means simple.

The practical consequence is that adoption is gated on readiness rather than belief. An agency drowning in administrative volume does not have the capacity to evaluate, configure and supervise a new system, however convinced it is. Which is why the useful question is not "does AI work" but "what is it allowed to do, who checks it, and who can turn it off" — set out in what you can automate and what still needs a licensed human.

Sources

Related reading