1 in 3 Independent Agency Employees Use AI – IA Magazine

Why “1 in 3” is a big deal for independent agencies

In the IA Magazine report on the “2025 Independent Agents at Work Study,” the headline number lands with real weight: one-third of agency
employees used AI for work in the past year. That’s not a “future trend” anymore; it’s a current workflow reality. The same report notes
57% of employees are interested in using AI for work, which suggests adoption pressure will keep risingespecially in roles where time
is already stretched thin.

Here’s the part that should make agency owners and managers sit up straighter: IA Magazine also highlights that only a small share of agencies
have formal AI guidancejust 12% have a well-defined AI usage policy. So while staff are experimenting, many are doing it without a
consistent playbook for privacy, accuracy, documentation, or client communications.

This gap creates a very “agency life” paradox: AI can help reduce workload, but unmanaged AI can introduce new errors & omissions exposure,
confidentiality risks, and compliance headaches. In a regulated industry built on trust, “we’ll figure it out later” can get expensive fast.

What AI use in an independent agency actually looks like

When people hear “AI,” they sometimes imagine robot underwriters or a chatbot selling umbrella policies at 2 a.m. Real agency use is usually
simpler and far more practical: small tasks that steal minutesrepeated hundreds of times a week.

1) Email drafting that doesn’t sound like a robot wrote it

Customer service reps and account managers often spend a surprising chunk of their day rewriting the same ideas: requesting loss runs,
confirming effective dates, explaining deductibles, outlining next steps for endorsements. AI can help draft a clean first version so the human
can add the final polish, correct details, and match the client’s tone.

Example: Instead of starting from a blank screen, an employee can generate a structured email that explains why a carrier needs updated payroll
estimatesthen double-check every number and remove any sensitive identifiers before sending.

2) Summaries of long, boring documents (the kind we all pretend we enjoy)

Policies, endorsements, proposals, risk surveysagencies swim in text. AI can provide a plain-English summary for internal use, highlight what
changed at renewal, or turn a complicated form into a short “here’s what to watch” note for a producer. This is especially helpful when
producers need the gist quickly but still want the source document for verification.

3) Marketing content that doesn’t require sacrificing your weekend

The IA Magazine coverage of Liberty Mutual/Safeco research has pointed to agencies seeing AI value in marketing supportlike brainstorming blog
ideas, drafting social captions, outlining newsletters, or creating campaign variations. Used well, AI becomes a content assistant, not a content
replacement: the agency still owns the strategy, compliance checks, and final voice.

4) Workflow automation and “next best action” nudges

AI is increasingly embedded in agency technologythink automation that suggests follow-up tasks, flags missing information, or helps match
clients with relevant products. Vendor tools can also surface cross-sell opportunities or identify service patterns that could be streamlined.

5) Training and knowledge support for newer employees

Younger staff are often more comfortable experimenting with AI tools at work, and research suggests that comfort correlates with overall
workplace adoption. AI can function like a searchable “first draft knowledge base” for internal questionsagain, with a human validating
anything that affects coverage advice or compliance.

Why AI is showing up now (and why it’s not just hype)

Agency adoption doesn’t happen in a vacuum. Across the broader workplace, AI use has risen quickly. For example, Gallup has reported that the
share of U.S. employees using AI at work at least a few times per year increased, and frequent use (a few times a week or more) has also grown.
Meanwhile, Microsoft’s Work Trend Index has emphasized that employees are bringing their own AI tools to workoften before leadership has a
formal plan.

That pattern maps neatly onto independent agencies: employees find tools that save time, then leaders scramble to build policy after adoption is
already underway. McKinsey has also noted that companies invest in AI but often struggle to reach maturitysuggesting the real challenge is
governance and leadership, not employee willingness.

For independent agencies, the driver is simple: clients expect speed and responsiveness, but agencies also face staffing pressure, workload
expansion, and the constant need to do more with the same (or fewer) people. AI, when used safely, promises a more efficient way to handle
repetitive work so humans can focus on relationship-building and complex problem-solving.

The not-fun part: risks that turn “helpful” into “uh-oh”

If AI were only upside, everyone would already have a policyand the IA Magazine report wouldn’t be calling attention to the policy gap.
Independent agencies handle sensitive personal and commercial data, so the risk profile is real.

Confidentiality and data exposure

Agencies may handle driver’s license numbers, loss histories, payroll data, health-adjacent details, and commercial financial information.
Feeding sensitive client info into the wrong tool (or the right tool with the wrong settings) is a recipe for privacy trouble.
A practical rule: if you wouldn’t paste it into a public website, don’t paste it into a public AI tool.

Accuracy, hallucinations, and “confident nonsense”

AI can produce plausible-sounding text that is still wrongespecially when asked to interpret coverage, compare policy language, or summarize a
document without careful checking. In insurance, the cost of a subtle error can be huge. AI output should be treated like a rough draft from an
intern who works fast and never sleepsand who must be supervised.

Bias and unfair outcomes

Regulators and standards bodies have repeatedly emphasized that AI systems can introduce or amplify bias. Insurance regulators have also been
paying close attention to how predictive models and AI affect consumers. Even if an independent agency isn’t building underwriting models, it
still interacts with carrier decisions, client advice, and marketing practicesareas where fairness and transparency matter.

Compliance and governance pressure is rising

The NAIC’s model bulletin on insurers’ use of AI has helped shape expectations around governance, risk management, and oversightespecially when
AI influences regulated insurance practices. On top of that, states such as Colorado have expanded AI-related governance obligations for insurers
using external consumer data and predictive models. Agencies should pay attention because carrier requirements and market conduct expectations
tend to flow downhill into distribution practices, vendor contracts, and documentation needs.

Marketing claims and client communications

If an agency uses AI tools to communicate, market, or explain coverage, it still needs to ensure the content is truthful, not misleading, and
appropriate for the client. U.S. regulators have warned more broadly about AI-driven deception and the risk of unsupported AI-related claims.
For agencies, the best defense is boring but effective: documented review steps and clear accountability for what goes out the door.

Build a practical AI policy (without writing a 47-page manifesto)

If only 12% of agencies have a well-defined AI policy, that’s an opportunity for leaders to get ahead. A strong agency AI policy isn’t a legal
treatiseit’s a set of simple, enforceable rules that protect clients and staff while still allowing productivity gains.

What a good agency AI policy should cover

• Approved tools: Which AI tools are allowed (and which are not), including browser-based chatbots vs. enterprise tools inside your agency systems.
• Data rules: What information is prohibited (PII, PHI-adjacent data, account numbers, claim details, etc.), plus redaction standards.
• Human review requirements: Which outputs must be checked by a licensed agent, account manager, or designated reviewer before use.
• Allowed use cases: Drafting, summarizing, brainstorming, internal templatesversus prohibited use (final coverage advice, binding decisions, or “send it as-is” communications).
• Documentation: When to save prompts/outputs in the file (for example, if it supports a client communication or a documented recommendation).
• Security and access: Password rules, MFA expectations, and how to handle AI add-ons or browser extensions.
• Training: A baseline training module so usage is consistent and safenot just “whoever clicked the right YouTube video first.”

A good policy should also include a simple principle that everyone can remember: AI can assist, but humans remain accountable.
That includes accountability for accuracy, compliance, and client trust.

A step-by-step rollout plan that actually fits agency life

You don’t need to “boil the ocean” (and if you try, your team will quietly update their résumés). The best agency AI rollouts start small and
scale only after the guardrails work in real workflows.

Step 1: Start with the “time thieves”

Identify repetitive work that consumes staff time but doesn’t require creative judgment: first-draft emails, meeting summaries, renewal
checklists, follow-up reminders, marketing outlines, or internal how-to notes.

Step 2: Create safe prompts and redaction habits

Build a short prompt library that reflects your agency’s tone, compliance expectations, and documentation style. Teach staff how to remove
identifying information and use placeholders like “[CLIENT NAME]” or “[POLICY NUMBER]” when testing drafts.

Step 3: Pilot with a small group and measure impact

Choose a small group of usersoften a mix of a producer, a CSR/account manager, and an operations/admin roleand define a 30-day pilot. Track:
time saved, error rates, rework, and client satisfaction impact. If time saved is real but error rates rise, adjust the process before expanding.

Step 4: Embed AI where it belongsinside secure workflows

Whenever possible, use AI features embedded in established agency platforms and vendor tools, rather than scattering random AI products across
desktops. Consolidation reduces the “shadow AI” problem and makes governance easier.

Step 5: Make training continuous (because tools keep changing)

AI tools evolve quickly. A quarterly refresher can keep staff aligned on data rules, client communication standards, and “what not to do”
exampleslike asking an AI to decide whether a claim will be covered. (Nope. Not today. Not ever.)

Where AI in independent agencies is headed next

If the past year was “experimentation,” the next phase is “integration.” Vendor ecosystems are developing AI-assisted workflows: better intake,
smarter routing, faster document handling, and improved client communications. Agent-focused resources are also mapping AI vendor landscapes to
help agencies understand what’s available and how to evaluate options.

At the same time, the standards conversation is getting louder: risk management frameworks, insurance governance expectations, and state-level
rules around data, models, and decision-making are shaping how AI must be managed. For agencies, the winners won’t be the ones who use the most
AI. They’ll be the ones who use AI responsiblywith clear policy, training, and human accountability.

Bottom line

“1 in 3 independent agency employees use AI” isn’t a gimmicky headlineit’s a signal that AI is already part of the independent agency
workflow. The bigger story may be the policy gap: tools are being used faster than agencies are defining the rules.

Agencies that respond with smart governanceclear do’s and don’ts, secure workflows, staff training, and real accountabilitycan turn AI into a
competitive advantage. Agencies that ignore it may still “use AI,” but in the least controlled way possible: quietly, inconsistently, and with
unnecessary risk.

Field Notes: Experiences agencies are having with AI right now (Extra )

The most common “AI experience” in independent agencies isn’t a dramatic system overhaulit’s a Tuesday afternoon problem that needs a Tuesday
afternoon solution. In many agencies, the first real win happens when a CSR uses AI to draft a renewal reminder email that’s polite, clear, and
doesn’t sound like it was written at 4:59 p.m. with one hand on the mouse and the other hand on car keys. The CSR still verifies the dates,
coverage references, and carrier requirementsbut the dreaded blank page is gone, and the tone stays consistent across the team.

Another frequent experience shows up in producer workflows. A producer prepping for a commercial account review might paste sanitized
notes into an AI tool to generate a meeting agenda: exposures to revisit, possible coverage gaps to discuss, and a checklist of renewal
questions. The result is not “AI advice”it’s a structured outline that helps the producer show up prepared. The producer’s expertise still
drives the conversation; AI just helps organize the mess of inputs into something usable.

Agencies also describe a learning curve that feels a lot like onboarding a new employee: you have to train it, supervise it, and correct it.
Early attempts can produce “confident nonsense,” especially if someone asks for policy interpretations in plain language without providing the
right context. The best teams respond by creating prompt guidelines and a simple internal rule: AI can summarize, but it can’t decide.
If the task involves coverage advice, binding authority, or anything that would matter in an E&O scenario, a licensed professional reviews
and owns the final output.

One of the most valuable experiences agencies report is discovering “shadow AI.” Leaders sometimes believe they haven’t adopted AIuntil they
learn employees are already using it for everyday drafting, brainstorming, and document clean-up. That realization can be uncomfortable, but it’s
also useful: it reveals what staff need help with most. And it often sparks the right next stepwriting a simple AI policy, not to punish people,
but to protect clients and the agency.

Finally, agencies frequently experience a shift in morale when AI is framed correctly. When leadership introduces AI as a way to reduce
busyworkrather than as a headcount replacementstaff buy-in improves. Teams get excited when they see AI saving ten minutes here, fifteen
minutes there, and turning those minutes into time for client conversations, proactive reviews, and training. The agencies that seem to get the
best results aren’t chasing “AI everywhere.” They’re building an “AI where it helps, humans where it matters” cultureand that’s a strategy
that fits the independent channel perfectly.