How employee benefits brokers are using AI

Seven in 10 benefits brokers now report having an AI strategy, according to Leader’s Edge reporting on the Council of Insurance Agents & Brokers 2026 employee benefits survey, which drew 166 brokers, account managers, practice leaders, and executives.¹ Most report being early in AI strategy development, with 55% of respondents planning limited to moderate investment, running from under $250,000 to $1 million a year.¹ The AI conversation has already reached clients, though. Six in 10 respondents say they’re actively discussing AI use cases in health benefit design with their clients.¹

What employee benefits brokers are using AI for

We see brokers start with three workflows. Two of them are places where a person is retyping something a document already says, and the third is answering questions those same documents already answer:

Carrier proposals

Working a case to market means collecting proposals, standardizing them into something comparable, and chasing whatever data was missing. Pasito’s Proposal Comparison puts AI agents on that work. They’re trained on documents from every major carrier in the U.S., they build a comparison table with your firm’s preferred fields already filled in, and AI checklists identify errors, inconsistencies, and missing data while there’s still time to chase it.

Guide and microsite production

Guide production repeats client by client, plan year by plan year, and language by language. With Pasito’s Benefits Guides, agents pull plan data from documents you already have, drop it into your firm’s templates with the client’s branding, and produce a guide in under 30 minutes, compared to about 30 hours by hand.

Benefits microsites are built separately, but they start from the same extracted data and the same eligibility logic. You pick a microsite template, the extracted plan data fills it, and you edit only what’s specific to that client. A few clicks later, there’s a working draft with auto-translated Spanish versions, so both language versions move on the same timeline instead of weeks apart.

A microsite is a place employees can receive benefits education, with articles, FAQs, and plan description pages built around the questions that come up during enrollment. It carries your firm's voice and the client's branding, and it runs on the same AI-extracted data and tokens that power the guides and the decision support plan comparisons. Pasito's structured token system means plan data shows up correctly on every microsite page, which is how a client's benefits education stays consistent across your whole book.

Employee questions and enrollment support

Employee questions arrive from the first announcement through the last day of the plan year. Whether those questions reach your team depends on your service model and capacity, and that varies widely between firms. When a firm does field questions, tools like Pasito’s AI Benefits Assistant answer employee inquiries that a client’s documents cover, around the clock, and the firm selects which of those documents the AI can draw from.

Decision support is the other piece brokers configure for their clients. It leads with a proactive, personalized benefits recommendation and removes the lengthy intake form, which is the friction point it’s designed around. A broker puts it in front of a client because it improves the experience that client gives their own employees, and that experience is part of what keeps the account.

Agentic AI in insurance: what a benefits broker can automate

Most AI people have used is generative. You prompt it, and it answers a question or makes an image. An agent works differently: you hand it a job with several steps, and it works through them in order and hands back finished work.

An agent does a task, and it can work repeatedly and reliably. Agentic AI takes a workflow with several steps in it, works through them in order, and hands back finished work, checking in only where it needs you.

In a brokerage, that looks like this: You drop in a stack of carrier proposals. The agent opens each one, finds the plan details wherever that particular carrier happens to put them, drops them into the data fields your firm uses, flags what’s missing, and lays out the comparison for you to review. Nobody typed anything, and nobody rebuilt the same table by hand for the fourth time this month. Pasito's agents are trained on documents from every major carrier in the U.S., which is what makes that first step hold up whatever format lands in your inbox.

The same pattern runs through the rest of the production layer, and it's all the same kind of work: retyping something a document already says. Agents pull core benefits data out of plan documents whether it arrives as a PDF, a CSV, or a Word file, build a branded guide in under 30 minutes, translate it into Spanish in one click, and build the microsite off that same extracted data. Our own figures put a benefits guide at about 40 hours by hand and a client's plan and rate comparisons at about 60 hours. That work compounds with how many clients you have, and none of it needs your judgment.

Talented brokers are needed more than ever

The judgment is what stays, and that's the design rather than an AI limitation. After all, you are the “advisor” in benefits advisor, and your strategic work is critical to client success. Our agents organize proposals for your review, but that review is the point. An agent can build the comparison and draft the guide before you open the file, and you're still the one deciding which options lead, what the recommendation says, and whether this is the year to push a plan design change. That's the work clients keep you for, and the only thing that changes after implementing AI agents is how many hours you have left for true advising.

Choosing which workflow to automate is half the equation. The other half is how you’ll implement it with a client: tools like Pasito’s benefits guides, microsites, and AI benefits assistant all run on your client's plan documents and have an interaction point with your client's employees. This typically raises questions from clients, in particular, how the automation will impact their employees, and what security measures are in place to adhere with compliance standards. 

What clients ask benefits brokers about AI and their data 

Some of what these tools touch is protected health information, which under HIPAA means individually identifiable health information held by a covered entity or a business associate. A tool that handles it will often have to clear a client’s security review, and larger clients will always ask. Five questions cover the ground that comes up most often:

Where does the data live, and how is it protected? A useful answer names the encryption standard in use at rest and in transit, and describes how one client’s data is kept separate from another’s.

Is client or employee data used to train models, and what does the model provider retain? Most benefits AI runs on a third-party LLM, which means data reaches that external ecosystem. So, it’s worthwhile to ask about the connection to the third-party infrastructure. Ask about data retention and model training to get a better picture of their AI security.

Where do the answers come from? An AI assistant drawing on general internet knowledge can describe insurance concepts with confidence, but it doesn’t know the plans your client offers. Ask whether AI responses are generated using the client’s own uploaded plan documents, and who controls which documents are in the agent’s knowledge base.

Who audited the controls? SOC 2 Type II is the primary attestation to ask about. Ask who performed the audit and what it covered.

A vendor built for benefits should have written answers ready. Ours are on Pasito’s security page.

How AI changes a benefits broker’s capacity

When brokers bring clients to market, that means more proposals to solicit from carriers and new benefit guides to create. Some of that work is worth it for large groups, but sometimes the math doesn’t make sense.

Unless your firm tracks service hours by account, those hours are spread across everyone’s week and are hard to account for since they aren’t sitting on a single line item. Want to see how AI-enabled workflows for proposal comparison and benefit guide creation can save your team time? Set the sliders below to match your own book, and you’ll see how many hours that work takes across a year.

Where to start with AI as an employee benefits broker

The easiest place to start using AI is with the work you already repeat: the same tasks for every client, every plan year, where someone is retyping what a document already says. That’s what agents handle well, and it is easy to check, so you find out quickly whether it holds up.

The judgment stays with you. Agents can organize the proposals, build the guide, and answer the questions your client's documents already answer, but which options lead and what the recommendation says is still your call. What changes is how many hours are left for that part.

Before any client's plan data moves, get the vendor's answers in writing. Every broker sells the same carriers, so service and technology are what separate firms now, and handling the diligence well is part of that. If you want to see how this runs on a real broker workflow, talk to our benefits experts.

FAQ

How are employee benefits brokers using AI?

The workflows brokers hand over first are production work: extracting plan data from carrier documents, comparing proposals, assembling branded guides and microsites, translating materials, and answering employee benefits questions. Leader’s Edge, reporting on the CIAB 2026 survey, found 70% of benefits brokers have an AI strategy, though 55% report a limited to moderate level of investment, in a band running from under $250,000 to $1 million a year.¹

Will AI replace employee benefits brokers?

Pasito has a firm stance here: no, AI should not replace advisors. Instead, what agents take over is monotonous production work like extraction, formatting, assembly, translation, and answering questions a plan document already covers. Renewal strategy, carrier negotiation, plan design, and more are still squarely in human hands. The opportunity lies in the recovered hours, giving every group the depth of attention they deserve

What is agentic AI in insurance?

Agentic AI completes a workflow end to end on its own. In benefits, an agent extracts plan data from carrier documents, standardizes it, builds comparisons, assembles guides, and routes anything it can’t verify to a person. Much of the writing about agentic AI in insurance describes property and casualty underwriting, where the workflows and the data differ substantially from group benefits.

What should a benefits broker ask an AI vendor about PHI and HIPAA?

Ask five things: the encryption standard and how client data is isolated, whether client or employee data trains the models and what the model provider retains, whether answers are grounded in the client’s own plan documents, who performed the security audit, and whether the vendor can explain all of it plainly. 

References

1. Leader’s Edge, “Employee Benefits Reach Inflection Point” (Council of Insurance Agents & Brokers 2026 employee benefits survey, 166 respondents). https://www.leadersedge.com/brokerage-ops/employee-benefits-reach-inflection-point
2. SHRM, “The Ultimate Guide to Hiring a Benefits Broker.” https://www.shrm.org/topics-tools/news/hr-magazine/ultimate-guide-to-hiring-benefits-broker
3. WTW, “Employers set to rapidly expand AI use in health and benefits, but execution gaps remain” (2026 AI Use in Health and Benefits Survey, 312 employers, 4.6 million U.S. employees, fielded January and February 2026). https://www.globenewswire.com/news-release/2026/05/19/3297669/0/en/Employers-set-to-rapidly-expand-AI-use-in-health-and-benefits-but-execution-gaps-remain-WTW-survey-finds.html
4. Prudential Financial, “Employers Embrace AI for Benefits, Employees Remain Cautious” (2026 Benefits & Beyond study, 760 employers and 3,096 employees, January 2026). https://news.prudential.com/us-en/latest-news/prudential-news/2026/q2/Employers-Embrace-AI-for-Benefits-Employees-Remain-Cautious