Your AMS Vendor Is Now an AI Platform

The shift that matters most for your agency is not happening at insurtech startups. It is happening inside the systems you already pay for.

Applied Systems began its AI journey in 2023 through the Applied AI Lab, and in 2024 acquired Planck, an insurance-specific AI company, to expand its AI capabilities across the policy lifecycle. Applied Book Builder, the risk intelligence tool inside Applied Epic for commercial cross-sell and upsell, is already in production at agencies running Epic today. The next several years will bring more of the same. AI surfacing inside features you already use, rather than as a separate product you have to evaluate and integrate.

Vertafore introduced the Velocity AI Platform at Accelerate 2026, a cloud-native AI stack integrating the new ReferenceConnect AI knowledge engine, six pre-built AI agents, and AgencyOne Certificates for automated certificate management. Whether you run Epic, AMS360, EZLynx, or something else, the platforms underneath you are being rebuilt around AI workflows.

This matters for one practical reason. Evaluating standalone AI tools without knowing what your AMS vendor is about to ship is how you end up paying for the same capability twice.

Before buying anything external, find two answers:

  1. Ask your Enterprise Admins and IT leads which features your AMS already supports that you haven’t enabled yet.
  2. Ask your AMS vendor’s account team what’s on the public roadmap.

The first conversation usually surfaces capabilities you forgot you had access to. The second one tells you what’s coming in the next few release cycles.

The AI Categories Worth Your Attention

The vendor landscape for insurance AI is enormous and fragmented. Agent for the Future’s directory tracks over a hundred AI-driven vendors across content generation, AMS providers, meeting note-takers, chatbots, marketing automation, sales assistance, knowledge bots, policy checking, retention, and video proposal tools.

For an independent P&C agency, only a few of these categories matter day to day. Here’s where your attention is well spent.

decor - person pointing at a laptop screen

AMS-Embedded AI

Applied Book Builder and upcoming vertical AI integrations for Epic users. For Vertafore users Velocity AI Platform and ReferenceConnect AI. These are the AI capabilities being built natively into your core system.

The pitch is that AI inside the AMS has access to your full book of business context that external tools lack. The reality is that this category is moving fast and what’s available today is different from what’s available six months from now. Treat AMS-embedded AI as the first thing to evaluate, not the last, because it shapes which third-party tools you would still need.

Policy Checking

Renewals are a volume problem in commercial lines. Every renewal needs to be compared against the prior term for coverage gaps, new exclusions, limit changes, and endorsement differences. Done manually, that’s a slow process that scales linearly with your book.

AI policy-checking tools compare current policies to renewals automatically and flag discrepancies. Agent for the Future lists over a dozen actively deployed policy-checking vendors including Quandri, FurtherAI, Exdion, Coverflow, and Patra. The category has visible adoption in the Applied Epic user community. At the same time, Applied has been working on policy-checking embedded inside Epic itself, which raises the buy-versus-wait question for agencies evaluating now.

The E&O angle is what makes this area serious. A coverage gap that slips through a manual review is an E&O exposure. A tool that catches it is reducing risk in addition to saving time. That is a stronger business case than pure efficiency.

There’s a caveat though. These tools depend on consistent, structured policy data in your AMS. If your Policy Types are inconsistent, if your specialty coverages live as PDF attachments without structured fields behind them, accuracy drops.

Policy checking AI does not fix bad data. It exposes it.

Meeting Note-Takers

The friendliest AI category, and the highest adoption inside agencies right now. Fathom, Fireflies, Otter.ai, Read, and Zoom's native transcription tools for recorded meetings, summarize them, and surface action items.

The value is straightforward. Your producers and account managers stop splitting attention between conversation and note-taking. They get accurate summaries they can paste into the activity log. They can share details with staff who weren’t on the call.

The risk is also straightforward. Get explicit permission before recording, or ensure with your legal team that your client agreements cover it. Consider establishing a policy on what gets transcribed and how long recordings are retained. Review the AI summaries before treating them as fact. Generative tools are confident even when they’re wrong.

Marketing Automation with AI Layers

Agency Revolution, ClientCircle, Applied Marketing Automation, Levitate, and similar tools have been around for years. The newer pitch is the AI segmentation layer on top. The promise is that AI will help find cross-selling opportunities, enable renewal outreach, and gauge client communications.

These tools are useful when an agency has both a real marketing function and clean book-of-business data to feed into them. They are not useful for agencies that are still doing renewal letters by hand, because the gap from manual to AI-personalized is too wide to bridge in one purchase.

Before buying more marketing automation, look at what you already have. Most agencies underuse the marketing tools – whether out-of-the-box basics, or upsold bundles – already integrated into Applied Epic or AMS360.

Knowledge Bots

ReferenceConnect AI, AlliBot, Linqura, The Intelligent Agent. These tools pull from documentation, policy language, and industry reference material to answer coverage and product questions for your staff.

The best use case is supporting newer producers and CSRs who don’t yet have the institutional knowledge to answer commercial coverage questions confidently. A knowledge bot trained on your agency’s standards plus industry reference material reduces the time a new hire spends seeking answers from senior staff.

The same caveat as everywhere else applies. The output is a starting point, not an answer. Your producer is still on the hook for what they tell the client. Having said that, this does not replace proper internal agency onboarding and training.

General-Purpose AI

ChatGPT, Claude, Gemini, NotebookLM, Perplexity. Agent for the Future identifies these as the most commonly used AI content generation tools for insurance agencies.

The good use cases are drafting, proofreading, summarizing long documents, rewriting emails to the right tone, and brainstorming responses to difficult client situations. The bad use cases are anything that involves putting client PII or carrier data into a public AI without understanding where that data goes. Remember, PII should not be getting into these tools, even more so if data sharing, “help improve data” settings, or general terms and conditions on data usage are unclear.

ACN’s CEO Brian Langerman recommends starting with identifying the top five repetitive weekly tasks your agency performs and testing whether prompts can automate any of them as a practical entry point. That’s the right approach. Small, contained experiments before any commitment.

decor - business people sitting by an office table with laptops

What to Watch For When Evaluating Any AI Tool

The principles below apply across every category above. They are also where the most expensive mistakes happen.

The Process Comes Before The Tool

If a workflow is broken, AI will scale the broken version. It will not fix it. Think about what specifically you are trying to change, what you are aiming to improve, and whether AI is even the right intervention.

Do you know how your business processes actually look in detail? Are they mapped? If the data structure in your Agency Management System is inconsistent, if your Policy Types are not standardized, if your Structure Groups do not reflect how the agency actually operates, no downstream AI tool will save you from those problems. It will read the inconsistent data and produce confident wrong outputs based on it.

This is the same point we made in our Applied Epic Optimization guide. Tools amplify your system’s configuration. They do not fix it.

There is a chance that what you need is workflow redesign, written procedures, training, and reliable auditing, not more technology taped on top. Standardization, training, or a configuration fix can solve problems that no AI tool will touch.

ROI Happens in Different Ways

Different tools create value in different ways. It is not always direct cost reduction.
Sometimes it’s reduced E&O exposure. Sometimes it’s freeing licensed account managers from administrative work so they can do client work. Sometimes it’s reducing the turnover that comes from staff doing repetitive low-value tasks. Make the business case in terms that reflect what’s actually changing, not in terms of made-up direct savings.

Internal Tools and Client-Facing Tools Have Different Bars

Treat these two categories as separate evaluations entirely. They have different risk profiles, different review needs, and different reasons to say yes or no.

An internal productivity tool that summarizes a meeting for your producer is low-stakes. If the summary is wrong, the producer corrects it before acting. A client-facing tool that drafts a coverage explanation, generates a certificate, or responds to a client question is high-stakes. Errors there carry E&O exposure and can damage the client relationship before anyone notices.

Here’s a useful way to think about it:

The further the AI output travels from the person reviewing it, the higher the bar. Internal-only AI gets a low bar. AI that touches client communication gets a much higher one.

Efficiency Is Not Effectiveness

AI can make some things faster. That doesn’t mean it’s making them better, and it doesn’t mean it isn’t harming your business somewhere else.

Efficiency is local. It measures isolated KPIs and ratios that may be misleading on their own. A chatbot that handles 80% of inbound service requests faster might be reducing renewal retention because clients feel they cannot reach a human when they need one. A policy-checking tool that processes renewals in minutes might be missing exception cases that a manual review would catch.

Effectiveness is the total net positive outcome of the changes you make.

Measure the right thing. Speed without quality is not an improvement.

Know When to Keep It Human

AI is good at analysis and information synthesis that would be overwhelming for humans. It can support decisions and serve as an interface for routine interactions. It is bad at showing empathy, judgment, and handling exceptions.

Your best clients call you when something has gone wrong. They want reassurance, the feeling of being heard, and someone who can navigate negative emotions with them. Putting a customer behind an AI agent wall during a claim or a difficult renewal may create resentment even when the metrics appear good. That resentment shows up later in retention numbers.

Decide deliberately where AI replaces human touch and where it supports it. Default to supporting, at least in areas where you can’t afford to fail.

Generative AI Needs Oversight

Generative AI does not have ultimate knowledge or the ability to reason. It produces plausible language by replicating structures it has learned. It is trying to “get it right enough”. Sometimes that’s enough. Sometimes it isn’t.

This matters most for any AI that touches client communication, coverage advice, or policy interpretation. Your producer is still legally and professionally responsible for what they tell a client. An AI draft is still just a draft.

ACN’s CEO, Brian Langerman, also frames three questions every agency should ask about any AI tool before deploying it:

  1. How is the data being used?
  2. Are the tools still following the rules?
  3. How are those tools using your data?

Those questions seem basic. They get skipped constantly.

How to Find and Evaluate AI for Your Insurance Agency

The short answer is that you do not need to scout the entire insurtech market. The agencies and tools that matter for your evaluation are concentrated in a few places.

The real answer is that the value of starting close to home (with your AMS vendor’s ecosystem and the user community around it) is that those tools have already been filtered for integration with the system you actually use. A best-in-class tool that doesn’t talk to your AMS might be worse than a good-enough tool that does.

Your AMS Vendor’s Partner Ecosystem

Applied Systems has a program for Certified Integration Partners. Vertafore has the Orange Partner Program. Both maintain directories of vetted integrations.

Working with a partner usually means easier data access, simpler integration, and the ability to operate with the core system in a meaningful way. Start there before evaluating standalone tools that haven’t been vetted by your vendor.

Ask your client success team for a partner recommendation. They know what other agencies your size are using.

The Applied Client Network and Connections

If you run Epic, ACN is one of the most useful resources you have. ACN’s Connections Publication regularly features practical AI guidance written by agency operations leaders, and the learning catalog hosts partner webinars on tools like Quandri, Outmarket AI, and Book Builder that go deep on actual Epic workflows.

These sessions cost nothing to access for members. They show real workflows from agencies using these tools, not vendor pitches. That’s a different signal quality than a demo.

Conferences That Are Worth Your Time

Applied Net, ACN Summits, or NetVU Accelerate are the events where AI vendors with serious Epic and Vertafore integration appear. ACN Summits run smaller events throughout the year in regional cities.

If a vendor is not showing at one of these events, that’s information. Either they’re early enough that the integration story isn’t built out yet, or they’re not committed to your core platform.

Industry-Specific Resources

Agent for the Future’s vendor landscape, maintained by Liberty Mutual’s agent programs team, is updated regularly and organizes AI-driven vendors by use case rather than by company. It’s a useful category-level reference when you’re trying to understand who plays in policy checking versus retention versus knowledge bots.

Insurance Consultants

You don’t need a big corporate consulting firm to evaluate AI. There are individual consultants and small firms with focused experience in independent agency operations.

The benefit is that a consultant who is not selling a specific tool may push back on whether you need technology at all, which is usually a healthier conversation than starting with a vendor demo. A consultant redirecting you away from technology to assess your workflows is a sign of responsibility.

The caveat is that some consultants might have incentives to recommend specific solutions. Ask directly whether they receive referral fees from any vendor before engaging.

Your Peer Network

Ask other agency operators what they’re using. An ex-colleague who moved on may already have solved your problem at their new shop. Other people in your market have most likely evaluated the same tools you’re considering. Pay attention to who is actually using a tool a year after they bought it, not just who is talking about it.

Industry groups extend this conversation further. Vendor community forums, Facebook groups, subreddits for insurance professionals readily respond and share tips. The honesty level is usually higher than what you’ll see in conference panels or vendor case studies, because nobody is on stage and nobody is selling. Search the archives before posting a question. Someone has usually asked the same thing already.

The Foundation That Decides Whether AI Works at Your Agency

A pattern runs through every category of AI for insurance agencies covered in this article. Policy checking depends on consistent policy data. Retention prediction depends on accurate book-of-business records. Knowledge bots depend on trained users who can spot a confident wrong answer. Marketing automation depends on segmented, accurate client data. AMS-embedded AI depends on the configuration of the AMS underneath it.

The short answer is that AI tools amplify whatever foundation they sit on.

The real answer is that most agencies have not done the foundation work, and they will not get the AI returns they’re hoping for until they do.

This part of the conversation is unglamorous. It does not get its own conference keynote. It is the work that decides whether you spend your AI budget well or waste it.

Four pieces of that foundation matter most, regardless of which AMS you run.

Your AMS Data Is Auditable and Audited

If your policy classifications are inconsistent across your book, if your specialty coverages live as PDF attachments without structured fields, if your carrier records are tangled between issuing entities, billing entities, and parent organizations, any AI tool reading from your AMS will produce outputs that reflect the inconsistency.

The specifics differ by system. The principle does not. For Epic users, we covered the audit framework in detail in the Applied Epic Audits guide. Auditing your data and workflows is the prerequisite for deploying AI on top of them.

Your AMS Configuration Is Actually Optimized

If your configuration is creating friction for your staff, an AI tool will scale that friction. If your structural hierarchy does not reflect how the agency operates, AI tools using that structure will produce reports nobody can act on.

Again, the process comes before the tool. Tools amplify configuration.

For Epic users specifically, we walked through this in the Applied Epic Optimization guide, but the same principles apply to other systems with their own terminology.

Your Staff Are Trained to Use AI Critically

Generative AI’s worst failure mode is the confidently-wrong output. That mode is only caught by humans who know enough to catch it.

New hires especially need foundational system training before AI tools enter their workflow, or they’ll learn to trust outputs they don’t yet have the experience to evaluate. Our Applied Epic Training guide covers the onboarding side of this for Epic users in detail. The same logic applies to any AMS-driven training program.

Your Admin Capacity Keeps Pace With Your Agency

This is the piece most often missed. Cleaning up data, normalizing carriers, restructuring hierarchies, reassigning workloads, integrating acquired agencies. This is the actual work that produces the clean data foundation AI tools need. Your Enterprise Admins are the people doing that work.

Most AMS platforms were built for one-at-a-time human operation. That serves your producers and account managers well. For your Enterprise Admins working across hundreds or thousands of records, that same design creates a volume bottleneck. Cleaning up inconsistent carrier records, aligning structures that touch hundreds of policies, correcting policy data across your book. Each of these is the groundwork AI initiatives sit on top of.

An AI tool sitting on top of inconsistent data will scale the inconsistency. The agencies that win the next wave are the ones that did the cleanup work first.

RecordLinker connects to Applied Epic and AMS360, giving your Enterprise Admins a bulk-friendly interface for the data foundation work. Your admins need a way to work on multiple records across entire books.

Cleaning up inconsistent carrier and entity settings, bulk employee creation and updates, structure management, policy and line-level corrections, and data quality checks happen in bulk. The changes get synced back to your AMS only once approved. What previously took days of admin time takes hours.

Our deepest admin integration is with Applied Epic, supported through our Certified Integration Partner relationship with Applied Systems.

Request a Demo

You might be evaluating AI tools for your insurance agency that need clean AMS data. You might already be buried in the bulk admin work that any AI initiative depends on. Either way, that’s exactly where RecordLinker comes in.

More Reading About Data and Insurance Operations

Visit these pages to learn more about agency operations, Applied Epic, and the data foundation agencies tend to miss: