68% of Agencies Are Expanding AI in 2026: Compliance Policy

An agency where producers quietly run AI chatbots and dialers with no sign-off shows why 68% of agencies expanding AI use in 2026 need a written compliance policy now. That policy, an AI tool inventory plus a sign-off and incident-response process, separates agencies regulators can trust from those left exposed.

How many agencies plan to expand AI use in 2026?

68% of independent agencies plan to increase AI use over the next 12 months, making 2026 the year AI shifts from experiment to default agency workflow. Yet only 8.29% of agencies currently use AI regularly and strategically, according to a 2026 CAIC governance playbook, leaving most agencies expanding tools faster than they can control them.

The jump is steep across the whole distribution chain. IndependentAgent.com's technology trends research found the share of independent agents using AI in at least one core workflow moved from 38% in 2024 to 64% in 2026, and Grant Thornton's 2026 AI Impact Survey puts the share of insurance agencies planning new AI investment at 98%. Carriers are moving in the same direction, though at different speeds by line of business:

Insurer line	AI use, plan, or explore rate (%)
Health	92
Auto	88
Home	70
Life	58

Life carriers report the lowest current AI usage among the four, at 58%, per coinlaw.io's 2026 industry statistics, which means life-focused agencies and IMOs are often adopting AI tools ahead of some of their own carrier partners rather than behind them.

What share of agencies have no written AI policy?

56% of agencies have no written AI policy or guidance in place, leaving more than half of the industry exposed while adoption accelerates. That gap is compounded by training: 44% of agencies rely on informal, peer-to-peer tech training instead of any formal enablement program, per the iDudes 2026 training-gap research.

The training gap widens the policy gap. When staff learn AI tools from each other rather than from a documented standard, the same chatbot or voice tool gets used differently by every producer, with no consistent rule on what client data it can see. ResourcePro's 2026 analysis of agency AI adoption frames this as a defining operational risk of the year: tools scaling faster than the rules that govern them. A written policy does not have to slow adoption; it has to catch up to where staff already are.

What AI risks worry insurance agencies most?

Data privacy and compliance risk is the top AI concern among insurance agencies, named by 24% of respondents as their biggest worry about AI tools. Inaccurate outputs rank second at 22%, per IndependentAgent.com's 2026 technology trends research, meaning nearly half of all agency hesitation traces to unverified answers and mishandled client data rather than price.

These worries are not irrational, but they are not the whole picture. Agencies also report clear upside: 60% cite operational efficiency and 52% cite staff productivity as their main motivations for adopting AI, per IndependentAgent.com's research, so the real question is not whether to use AI but how to use it without opening a compliance gap. A governance policy resolves both sides of that tension: it names which tools may touch client PII, and it gives producers a documented answer when a client or examiner asks how an AI-assisted call or quote was handled. Kadence, for example, ties its Voice AI answering and follow-up to logged consent and National Do-Not-Call suppression on every outbound call and text, so the compliance rule lives inside the workflow itself instead of sitting in a binder no one opens.

What belongs in an agency AI governance policy?

An agency AI governance policy needs five elements: a full AI tool inventory, a named approval process, acceptable-use rules, vendor diligence, and incident response. Per the CAIC's 2026 governance playbook, it should require sign-off before any new tool touches client PII, and ban confidential data in unapproved tools.

Each element does a specific job:

• Tool inventory: lists every AI tool in use, who uses it, which workflow it touches, and what client or premium data flows through it.
• Approval process: names one owner who signs off before any new AI tool works with client data or premium information.
• Acceptable-use rules: define what AI may draft or suggest, and require a licensed producer to review AI-generated work before it reaches a client.
• Vendor diligence: documents how each AI vendor handles data security, since liability for that vendor's tool does not shift away from the agency.
• Incident response: sets a written protocol for AI-related errors or data leaks, including who is notified and how fast.

Because the tool inventory is the hardest line item to keep current, agencies that consolidate lead capture, calling, and follow-up into one system have an easier time proving what data an AI tool actually touched. Kadence is AI built to grow life insurance distribution, front to back office, and its CRM keeps every inbound lead and AI-handled interaction in one pipeline, which turns an agency's tool inventory into a live record instead of a spreadsheet nobody updates.

How fast can an agency build an AI governance program?

An agency can draft a first-version AI governance policy in about two weeks and build a fully defensible program from zero within 90 days. The CAIC's 2026 playbook frames those two milestones as sequential: a written policy first, then documented vendor diligence, audit trails, and incident response layered in over the following weeks.

A workable build sequence looks like this:

1. Weeks 1 to 2: draft a policy covering tool inventory, approval sign-off, and acceptable use, then get written acknowledgment from ownership or senior management.
2. Weeks 3 to 6: run vendor diligence on every AI tool already in use, documenting each vendor's data-handling and security posture.
3. Weeks 7 to 10: build the incident-response protocol and start an audit trail for AI-assisted client interactions.
4. Weeks 11 to 13: pilot the policy on one workflow, then extend it agency-wide by day 90.

This timeline is an operational target, not legal advice; agencies operating in multiple states should confirm specifics with counsel given how unevenly state adoption of AI guidance is unfolding.

How are carriers using AI in 2026?

Carriers are moving fast on AI, with 88% of auto insurers, 70% of home insurers, 58% of life insurers, and 92% of health insurers using, planning, or exploring AI in operations. Per coinlaw.io's 2026 industry statistics, that pace is pushing carrier documentation and vendor-approval standards onto the agencies they appoint.

Wolters Kluwer's analysis of AI in insurance regulation describes oversight moving "from principles to operational accountability," and that shift shows up first at the carrier-agency interface: appointment questionnaires, marketing-material approvals, and vendor lists increasingly ask what AI tools an agency runs and how it controls them. Life carriers report the lowest current AI usage of the four lines, at 58%, which means a life-focused agency or IMO network that documents its own AI governance now can look more disciplined than the carrier partner reviewing its appointment file.

What is the gap between AI adoption and governance?

The gap between AI adoption and AI governance is wide: 68% of insurance companies say their AI controls are fragmented across teams and tools. Only 24% are fully confident in those controls, and Grant Thornton's 2026 AI Impact Survey found 56% cite regulatory uncertainty as a top barrier to scaling AI.

That fragmentation persists even as maturity climbs: 62% of insurance companies rate their AI maturity as scaling across functions, per the same survey, meaning adoption is outrunning control at the carrier level too, not just among agencies. The lesson for distribution is direct: a single owner, a single inventory, and a single incident-response step close more of that gap than adding another AI tool ever will. Agencies that treat speed to lead and governance as one system rather than two competing priorities avoid inheriting the same fragmentation their carrier partners are now working to fix.

Why is a written AI policy a competitive edge?

A written AI policy becomes a competitive edge because it lets an agency prove disciplined AI use to carriers, regulators, and clients while most competitors cannot. With 56% of agencies lacking a written AI policy, per IndependentAgent.com's research, the other half stand out fast during carrier appointment reviews or E&O audits.

Carriers increasingly ask appointed agencies to show, not just claim, how they handle AI-assisted outreach. An agency that can produce a one-page policy naming its tools, its approval owner, and its consent-and-opt-out rule for every outbound AI call moves through appointment and audit conversations that stall competitors relying on ad hoc AI use. Kadence's Voice AI, for instance, records consent and honors National Do-Not-Call suppression on every automated call and text it places, so the policy's outbound-consent line is backed by a system record rather than a promise, and an AEO-built site that gets an agency cited directly in AI search results signals that same operational discipline to prospects before they ever call.

What regulatory rules shape agency AI governance?

The NAIC's AI Model Bulletin shapes agency governance the most, and 23 states plus Washington, D.C. had adopted it by late 2025. Per the NAIC, it does not create new law but codifies expectations for transparency, accountability, fairness, privacy, and safety under existing unfair-trade-practices law.

By early 2026, per JD Supra's analysis of AI governance expectations, over half of all U.S. states had adopted the Model Bulletin or similar guidance, and state enforcement mechanisms apply fully to AI-assisted decisions in marketing, underwriting, pricing, and claims. The bulletin is principle-based, so it does not mandate one testing method; it requires insurers, and by extension the agencies acting under their authority, to prove validation against these five expectations:

NAIC principle	What it expects from an AI system
Transparency	Clear documentation of what the tool does and what data it uses
Accountability	A named owner responsible for the tool's outputs
Fairness and Equity	Evidence the tool does not produce biased outcomes
Privacy	Controls limiting exposure of client and premium data
Safety	Testing and monitoring for errors before and after deployment

An agency AI governance policy that maps directly to these five principles is easier to defend in an examination than one written from scratch around internal preferences alone.

Does an agency stay liable for a vendor's AI tool?

Yes, an agency stays fully liable for any AI tool it uses even when a third-party vendor built it, because liability does not transfer to the vendor under current regulatory expectations. A 12-state pilot of the NAIC's AI Systems Evaluation Tool launched in early 2026 to formalize how examiners test that accountability.

This is exactly why vendor diligence sits inside the governance policy rather than outside it. An agency that can document why it chose a given AI vendor, what data the vendor can access, and how the vendor handles a breach holds a materially stronger position in an examination than an agency that simply assumes the vendor is responsible. The same documentation discipline extends to the back office: Kadence's back-office commission tracking keeps the money side of the book, who wrote what and what a carrier paid, in one visible record, which is the kind of audit-ready trail examiners and carriers increasingly expect across both sales and servicing workflows.

Agency policy element	NAIC principle it satisfies
Tool inventory	Transparency
Approval sign-off	Accountability
Acceptable-use rules	Fairness and Equity
Vendor diligence	Privacy
Incident response	Safety

Ready to see how a governed, faster front office actually runs day to day? Book a demo and walk through the workflow with the Kadence team.

Sources

• AI Governance for Insurance Agencies: A 2026 Playbook
• Insurance insights: 2026 AI Impact Survey | Grant Thornton
• AI for Insurance Agents in 2026: Adoption Hit 64%, The Industry ...
• Technology Trends Report - IndependentAgent.com
• The NAIC Just Added AI Governance to Your Insurance ...
• Why AI in insurance agencies is defining 2026
• 2026: The Year AI Goes Operational in Insurance
• AI in Insurance Industry Statistics 2026: Shocking Growth ...