About This Report

The independent insurance agent has been declared obsolete more than once. This report looks past the predictions and asks a narrower question: what does the evidence actually say about the independent distribution channel, and which parts of the agent's job still create value as technology improves?

The report draws on government data, insurance-regulatory information, industry research and original market studies available through August 2026. Sources are collected in Appendix B so the body of the report can be read as a report rather than a string of citations. Where the evidence supports more than one interpretation, that is stated plainly.

This is not an argument that every agent is valuable, that every agency is healthy, or that artificial intelligence cannot replace meaningful parts of insurance work. In several places, the evidence points in the opposite direction.

Executive Summary

The insurance agent has supposedly been on the verge of extinction for a long time.

The internet was going to do it. Then online quoting. Then direct-to-consumer insurance, Insurtech, Wall Street money and now it is artificial intelligence.

At some point, when the same prediction keeps getting made, it makes sense to stop arguing about it and look at what actually happened.

I went into this expecting to make the argument that people will eventually reject AI, and ultimately crave human interaction. I expected that to be my argument. But I left with something else.

Independent insurance agencies still placed about 62% of all property and casualty premium written in the United States in 2025. They accounted for 87.7% of commercial-lines premium and 39.5% of personal-lines premium. Personal-lines share is actually higher than it was four years earlier, when independent agencies held about 36.7% of that market. [1]

Those numbers do not describe a distribution system that has been pushed aside by technology. But I would not take too much comfort from them either.

Something is changing. And I think we have spent too much time asking whether insurance agents are going to disappear, when the better question is what insurance agents are actually going to be paid to do.

Artificial intelligence can already summarize underwriting guides, compare policies, organize notes, research markets, write routine communication and work through information much faster than a person can. A 2026 study of 1,149 independent-agency principals and employees found that 65% had used AI for work during the previous year. A year earlier, that number was only 37%. Among agents using AI, the reported average time savings was about four hours per week. [2]

Consumers are changing too. Digital purchasing is now a normal part of the insurance experience, especially in simpler personal-lines transactions. That creates a very real case against the traditional insurance agent.

A customer does not need an agent to download an ID card. They increasingly do not need one to obtain a quote. They may not need one to explain basic policy terminology. Soon, they may not need one to compare fairly complicated insurance products. Those parts of the job are vulnerable.

At the same time, something almost contradictory is happening. Insurance itself is not necessarily getting easier.

The homeowners market has become more difficult in many parts of the country. Between 2018 and 2024, inflation-adjusted average homeowners premiums increased between 18.3% and 43.3%, depending on the region. Company-initiated nonrenewal rates increased between 96% and 216%. [3]

Medicare Advantage now enrolls 35.2 million people, or about 55% of eligible Medicare beneficiaries. Yet the years of broad, rapid expansion appear to be changing. Most of the net enrollment growth from 2025 to 2026 came from Special Needs Plans. [4]

The Bureau of Labor Statistics does not expect insurance sales agents to disappear over the next decade either. It projects employment to rise about 3% from 2025 through 2035 and estimates roughly 43,100 openings per year. Many of those openings are replacements rather than entirely new jobs. [5]

None of this proves that insurance agents are safe. It tells us the story is more complicated.

Technology is making it cheaper and faster to find information and handle routine administrative work. That means simply having the information isn't worth what it used to be. Understanding what the information means, however, may actually become more valuable. That is what this report is about. Not whether technology is good or bad, or even if agents deserve to survive. And not even if the insurance business was better twenty years ago.

The question is: What does an independent insurance agent still do that has real value, and how much of that value can technology eventually replace?

The answer is going to determine which agencies thrive over the next decade.

The Independent Agent Was Supposed To Be Gone By Now

There is a tendency in technology to assume that removing friction automatically removes the intermediary. Sometimes it does.

Insurance looks vulnerable to the same logic. Rates can be displayed online. Applications can be completed online. Policy documents can be delivered electronically. Consumers can communicate directly with insurance companies. So why did independent insurance distribution not collapse?

Commercial insurance provides one answer. Independent agencies still place almost 88% of commercial-lines written premium in the United States. [1]

That number makes more sense when you stop thinking about insurance as one product. A restaurant needs insurance. So does a dentist, a roofer and a technology company. But, that is about where the similarities end.

A restaurant can have property, general liability, workers compensation and food-related exposures. A roofing contractor brings height exposure, subcontractors, commercial vehicles and a liability problem many carriers do not want. A dentist may need malpractice, cyber, employment practices and business-property coverage.

Then the carrier matters. Carrier A may like restaurants. Carrier B may not. Carrier C may write contractors but will not touch roofers. Carrier D may write them if subcontracting stays under a certain percentage. One carrier may like a risk in one state and refuse the same business somewhere else. And all of that can change.

Commercial insurance never became a perfectly clean marketplace where every risk fits neatly into a standardized box. The real job is often not, 'Who has the cheapest quote?' It is, 'Who will write this correctly?' That is a much harder problem to automate.

Personal lines should be more vulnerable. A relatively ordinary auto policy is far easier to standardize than a commercial account. Homeowners insurance is familiar. Direct insurers have been advertising to consumers for decades. Digital distribution is well established.

Yet independent-agent personal-lines market share increased from 36.7% in 2021 to 39.5% in 2025. [1] That surprised me.

But we need to be careful with it. It does not necessarily mean consumers suddenly decided they love independent insurance agents. The insurance market itself became difficult during that period. Rates went up. Underwriting tightened. Companies became more selective.

When the market is easy, having one insurance company may be enough. When the market becomes difficult, having several companies becomes more useful. That may explain at least part of what happened.

And if that is true, it tells us something important about independent distribution. The independent agent may become more valuable when the insurance market becomes less predictable.

The Independent Agency Is Still A Good Business

There is another way to test whether this distribution model is actually healthy. Follow the money.

The Big 'I' and Reagan Consulting Best Practices Study tracks selected high-performing independent agencies. The 2026 update showed continued profitability even as organic growth slowed. [6] These are intentionally strong agencies, so the results should not be treated as the economics of the average insurance office.

But if the basic independent-agency model were already broken, we probably would not see high-performing firms continuing to produce strong operating results.

There is a catch. Growth slowed. And I think that part deserves more attention than it usually gets.

P&C agencies have spent several years receiving help from the hard market. Suppose a commercial account has a $100,000 annual premium. The carrier raises it to $115,000. Assuming the commission stays similar, agency revenue goes up. The agency did not find another customer. Nobody made another sale. The insurance simply became more expensive.

That matters when people look at agency growth statistics. If rate increases slow, the agencies that truly know how to generate new business should begin separating themselves from agencies that were getting dragged upward by premium inflation.

And as the easy growth gets harder. Agencies will have to create growth the old-fashioned way. Productivity becomes more important. So does producer development, technology and the ability to understand the markets better than your competitors.

What Happens When The Experienced Agents Leave?

The Bureau of Labor Statistics projects about 43,100 openings per year for insurance sales agents between 2025 and 2035. Employment itself is projected to grow about 3%. Many openings will come from people retiring or leaving the occupation rather than entirely new positions. [5]

That distinction matters. Insurance has an enormous amount of knowledge that was never properly documented. Some of it is written down. A lot of it isn't.

You can teach a new agent how to run a quote. You can teach the application. You can explain deductibles and coinsurance. You can give someone an underwriting guide. Experience is different.

An experienced P&C agent learns that a carrier technically writes a class but never seems to be competitive on it. They know which underwriter is worth calling. They recognize when carrier appetite has started to shift before the formal bulletin reaches the agency.

A life insurance agent knows that the prescription list may tell a completely different story than the client's answer to the health question. A Medicare agent knows that the plan that looks best on paper may not work if the provider directory is wrong. A Medigap agent learns that today's cheapest rate does not always tell you much about what the customer will be paying five years from now.

Technology can help capture some of that knowledge. It probably should have happened sooner. Because every time an experienced agent retires, some of what they know walks out of the building with them.

AI may end up being part of the solution to that problem. Which is one reason the 'AI versus agents' conversation is too simple.

Then The Homeowners Market Broke

One of the strongest arguments for independent distribution can be seen in what happened to homeowners insurance.

The NAIC's national analysis of 2018 through 2024 data found that inflation-adjusted average premium per homeowners policy rose in every region of the country. The increase ranged from 18.3% to 43.3% depending on the region. At the same time, company-initiated nonrenewal rates increased between 96% and 216%. [3]

That is not just an insurance-industry statistic. Put yourself in the customer's position. You bought your house. You bought the insurance. You pay the bill every year. Then the premium jumps. Maybe it jumps again. Then you receive a letter saying the insurance company no longer wants the house. Now what?

This should be the perfect independent-agent situation. The customer has a problem. The agent has multiple insurance companies. Find another one.

Except sometimes the agent cannot. Maybe the roof is too old. Maybe the ZIP code is restricted. Maybe replacement cost is too high. Maybe the carrier has too much exposure nearby. Maybe eight of the ten companies the agent represents have all tightened underwriting.

This is where the independent-agency model runs into a problem that I do not think gets enough attention: access to carriers is not the same thing as understanding those carriers.

Carrier appetite is constantly moving. And somehow, in 2026, an incredible amount of that information is still spread through emails, portal notices, underwriter phone calls, wholesalers, carrier reps, agency meetings and agents talking to other agents.

Sometimes the agent learns the underwriting changed after submitting the application. That is an information problem that should have been solved a long time ago.

The same Big 'I' data showing independent-agent market-share strength also shows agencies increasingly using specialty markets. Surplus-lines utilization reached 9.9% in 2025, above the five-year average. Private flood utilization reached 52.6%, also above its recent average. [1]

The market is giving agents more places to put difficult risks. But every additional market brings another set of rules. That creates more choice. It also creates more things to understand.

Something Else Is Happening To The Independent Agency

While everyone debates whether consumers still need agents, a quieter change is taking place behind the scenes. Agencies are being bought. A lot of them.

Industry M&A tracking shows hundreds of announced U.S. insurance brokerage transactions each year, with private-capital-backed buyers responsible for a large share of activity. [7]

This creates an interesting contradiction. On one hand, independent insurance agents are supposedly an outdated distribution model waiting for technology to finish them off. On the other hand, sophisticated investors keep spending large amounts of money buying insurance distribution.

They are probably not doing it for nostalgia. Insurance agencies have characteristics financial buyers like: renewal revenue, client retention, recurring demand, relatively predictable cash flow and opportunities to combine firms.

The more interesting question may therefore not be whether independent distribution survives. It may be whether independent ownership survives at the same rate.

The office down the street may still have the same name ten years from now. The producer may be the same person. The customer may not notice anything changed. But several layers above that office, the actual owner could be a large brokerage backed by institutional capital.

There are real advantages to scale. A large organization can have dedicated cybersecurity, compliance, IT, data analysts, marketing, training, specialists and centralized service. The five-person local agency cannot build all of that internally.

But the five-person agency has advantages too. It can make a decision quickly. It can specialize. It can know the local market better than someone hundreds of miles away. It can operate without layers of bureaucracy.

I do not think small agencies disappear. I do think 'small and average' becomes a harder place to live. The smaller agency probably has to be unusually good at something.

Medicare Advantage: What Happens After Twenty Years Of Growth?

Medicare Advantage may be the best example of why insurance agents increasingly need to understand the market behind the product.

In 2007, approximately 19% of eligible Medicare beneficiaries were enrolled in Medicare Advantage. By 2026, it was 55% - about 35.2 million people out of 64.2 million beneficiaries with Parts A and B. [4]

That is one of the biggest shifts in health-insurance distribution in the country over the last twenty years. Medicare Advantage was once the alternative. Now it is the majority.

That growth created enormous opportunity for carriers, FMOs and agents. But the growth story is changing.

Special Needs Plans now account for approximately 23% of all Medicare Advantage enrollment. About 85% of the net increase in MA enrollment from 2025 to 2026 came from SNPs. C-SNP enrollment increased by 45%. [4]

I think that matters more than the headline growth number. The opportunity is not disappearing. It is moving.

A market dominated for years by broad expansion is becoming more specialized. That affects the agent. It affects training. It affects marketing. It affects the types of consumers becoming the growth opportunity.

There is also concentration. UnitedHealth Group and Humana together account for about 46% of national Medicare Advantage enrollment. UnitedHealth's share fell to about 26% in 2026 while Humana increased to about 20%. [4]

Those are large movements. But even they can mislead the local agent. Because Medicare Advantage is not really one national market. Plans are local. Networks are local. Benefits are local. Competition is local. Carrier strength is local.

A carrier losing enrollment nationally can still be doing very well in a particular county. That makes a different set of questions more useful: Which carrier is gaining share here? Has MA penetration in the county continued rising? Has it flattened? Are SNPs driving growth? Are members entering Medicare Advantage, or are existing MA members simply moving between plans?

National data gives you the big picture. Local data tells you where the business is.

Medicare Is Complicated. That Does Not Automatically Make The Agent Valuable.

A Medicare beneficiary may be trying to understand Original Medicare, Medicare Advantage, Medigap, Part D, provider networks, formularies, prior authorization, premiums, deductibles, maximum out-of-pocket exposure, supplemental benefits, medical underwriting, guaranteed-issue rights and enrollment periods.

That creates room for a knowledgeable agent. It also creates room for a bad agent to sound knowledgeable. Those are not the same thing.

Complexity itself is not a value proposition. Helping the customer work through the complexity is.

And AI will eventually become very good at parts of that process. Imagine a Medicare system with access to a beneficiary's doctors, medications, hospitals, claims history, travel patterns and budget. It could theoretically compare every plan in the county at the same time. A human cannot manually process that much information as quickly.

So the traditional defense - 'Medicare is complicated, therefore consumers need agents' - is not enough.

The better question is what happens after the computer produces the answer. Is the provider directory accurate? Does the client actually understand the network restriction? They said they do not travel much. Then you discover they spend eight weeks with their daughter in another state every summer. Can the customer get a Medigap policy later if they change their mind? What happens if the current plan changes substantially next year?

The machine will get better at those questions too. But that is the level at which the future agent needs to operate. Not reading copays off a screen.

We Should Also Talk About The Commission

Agents sometimes become defensive when compensation comes up. They shouldn't. Insurance professionals get paid. There is nothing unusual about that.

But commissions create incentives. Different products may pay differently. Different transactions may pay differently. That creates a potential conflict of interest.

It does not mean the advice is bad. But if the advice is good, it should be able to survive transparency.

Why did you recommend Carrier A? Why not Carrier B? Why Medicare Advantage instead of Medigap? Why replace the policy? What is the customer giving up? What are they getting?

A good recommendation should still make sense after the client understands how the agent gets paid. I think this will become more important as consumers gain access to better comparison technology.

And frankly, that is probably healthy for the industry. Better transparency should hurt weak advisors more than good ones.

Medigap: The Product Looks Simple. The Market Isn't.

Medicare Supplement looks straightforward because the benefits are standardized. A Plan G from one company covers the same standardized benefits as a Plan G from another. So a consumer naturally looks at price. And price absolutely matters.

But today's price is today's price. What happens afterward? This is where the carrier starts to matter.

Rate history matters. Claims experience matters. The size and age of the block matters. The state matters. Underwriting matters. Switching rights matter.

A healthy 65-year-old may be able to move between carriers fairly easily. That same person at 73 with several health conditions may not. That means a decision that looks reversible today may not be nearly as reversible later.

This is why the question, 'Who has the cheapest Plan G?' is incomplete.

The better question is: 'Who has the best Plan G for this person, in this state, at this point in time, considering what we know about the carrier and the customer's ability to change later?'

The state-level differences matter too. A company can be a major Medigap carrier in one state and almost irrelevant in another. Rate patterns differ. Underwriting differs. Consumer protections differ.

Again, we end up back at the same place. The data exists. The useful part is interpreting it locally.

Final Expense Is A Great Example Of Where Technology Helps The Agent Before It Replaces The Agent

Final Expense looks like exactly the type of insurance that should lend itself to direct distribution. Small face amounts. Simplified issue. Usually no exam. Fairly simple consumer proposition.

But the underwriting can become complicated quickly.

Take diabetes. That is not enough information. Is the person taking insulin? For how long? Neuropathy? Kidney issues? Any amputations? Then look at the prescriptions. Eliquis. Why? Atrial fibrillation? Blood clot? History of stroke? Add COPD. Or oxygen. Or congestive heart failure. Now the answer changes again.

One company may offer immediate coverage. Another offers graded benefits. Another will not write the applicant. The product is simple. The placement is not.

And this is one place where technology should improve the agent's job very quickly. There is no good reason a producer should be manually searching ten PDF underwriting guides to answer the same question. Software should do that. AI should make it even faster.

That does not remove the agent. It removes some of the inefficient work around the agent.

The agent who relied on memorizing underwriting manuals loses an advantage. The agent who understands underwriting and now has better tools becomes more capable. That difference is going to show up throughout the industry.

AI Is Not Coming. It Is Here.

The speed of AI adoption inside independent agencies may be one of the more important findings in this report.

Usage increased from 37% to 65% in one year among agents surveyed by Liberty Mutual. Weekly use increased from 18% to 41%. Agents using AI reported saving an average of four hours per week. [2]

Four hours matters. An agency with ten employees saving four hours per person effectively creates another 40 hours of weekly capacity. That is another workweek.

And these are still fairly early uses: meeting notes, marketing content, policy comparisons, routine tasks and document handling. The technology is nowhere close to finished.

This is why I think the first major AI disruption in insurance is being misunderstood. Everybody wants to know when AI replaces the agent.

The more immediate question is: What happens when one agent starts using AI against another agent who doesn't? That competition has already started.

Two Agencies Walk Into The Same Market

Imagine two agencies. Same town. Similar carrier appointments. Similar products.

Agency A works more or less the way it did ten years ago. Underwriting knowledge is stored in PDFs and people's heads. Someone manually writes every follow-up. Carrier updates sit in email inboxes. Market research happens when somebody has time. Employees spend a lot of time finding information.

Agency B uses technology aggressively. The underwriting rules are searchable. AI summarizes carrier updates. Routine follow-up is automated. Customer meetings are summarized. Market data is organized. Employees spend less time searching and more time using the information.

Agency B does not need to eliminate jobs. It only needs to become 20% more productive. That is enough to disrupt the market in that town.

Twenty percent more capacity, producer time and lower service cost. The agency becomes more profitable and more valuable. You don’t have to imagine an industry run by AI to see disruption, 20% for production is enough to change the economics of any business.

That is disruption.

The Strongest Case Against The Independent Agent

I do not want this report to become an argument written by an insurance agent explaining why insurance agents are the best thing since sliced bread. There are very good reasons some parts of this profession could disappear; if all I'm doing is taking information a customer could find themselves and repeating it back to them, I'm not sure that's much of a value proposition.

Direct distribution works – that’s an established fact. If a customer understands the product and can complete the entire transaction in ten minutes, adding another person to the process may not accomplish much.

AI is also attacking what was traditionally one of the agent's stronger arguments: complexity. We used to be able to say insurance is complicated, therefore people need someone who understands it. But AI is very good at complicated information. It can read more than a human. It can compare more. It can do it faster.

A human recommendation is not automatically superior either. Agents make mistakes. Agents have favorite carriers. Agents sometimes sell what they know instead of what is best. Agents can miss exclusions. Agents can misunderstand underwriting. Compensation can influence behavior.

So 'trust the human' is not enough. The agent still has to earn the trust.

Then there is the administrative side of the agency: data entry, scheduling, document summaries, basic service, routine email, renewal preparation and simple research. A lot of this is going to require fewer human hours.

I do not think half the industry wakes up unemployed one morning. It will happen much more quietly. A book that needed ten people may need nine. Then eight. The book gets larger. Revenue per employee rises. Nobody announces that AI replaced two jobs. But the economics changed anyway.

And finally, the digital experience is going to improve. Agents should not compare themselves against a frustrating carrier website in 2026 and assume that is what digital insurance will look like ten years from now. It will not.

The systems will get much better. The independent agent has to get better too.

What Consumers Want Is Probably More Complicated Than Digital Or Human

I don't think most consumers wake up thinking, 'I am a digital insurance consumer,' or, 'I believe in traditional agent distribution.' They just want the easiest reasonable way to solve the problem.

If they need an ID card, they want the app. If they need to change their address, they probably don't want to call the office. If they need a routine document, let them download it.

Now ask them whether they want to decide how to insure a $4 million business the same way. Different situation.

A simple auto policy for a low-risk driver is not the same decision as insuring a contractor, selecting Medicare coverage for someone with several specialists and expensive medications, placing Final Expense coverage on someone with several medical conditions, or finding homeowners coverage after a nonrenewal.

This is why I do not think 'Will consumers use agents?' is a very useful question. But, for what? That is the better question.

The agent is weakest when the product is standardized, underwriting is straightforward, options can be easily compared, the transaction is low consequence and little advice is needed afterward. This is the case with many insurance products. But, the agent becomes more useful as those things move in the other direction – when there is a real difference between what company A and company B offer, and underwriting is not straightforward. This is not really an argument for human beings. It is an argument for expertise.

Today, that expertise often comes from a human. In the future, it may increasingly be a combination of human experience and machine intelligence.

Insurance Has An Information Problem That Should Embarrass Us A Little

There is an absurd amount of insurance information. CMS datasets. NAIC information. State filings. Carrier bulletins. Underwriting guides. Rate notices. FMOs. Wholesalers. Carrier portals. Webinars. Emails. Spreadsheets. Industry publications. Forums. Social media.

And after all of that, an insurance agent can still struggle to answer: Who wants this risk right now?

That is strange.

A Medicare agent can see every plan available in a county and still have no idea which company has been taking market share. A Medigap agent can produce a rate in seconds but not know whether the carrier has been gaining or losing insured lives. A Final Expense agent may have fifteen underwriting guides and still need to manually compare all of them. A P&C agent may have fifteen carrier appointments and send an account to three companies that never wanted it.

The industry has become very good at producing information. It has not become nearly as good at organizing the information into something an agent can use.

That may be one of the largest opportunities in insurance technology.

This Is Where Market Intelligence Starts To Matter

There is a difference between data and intelligence.

Let’s say Medicare Advantage penetration in a certain county is 61%. That is data.

Now suppose it was 41% five years earlier. That tells us more.

Now, suppose Carrier A gained ten percentage points of market share while Carrier B lost six percent points. Now we have something worth looking into. Now, we potentially have market intelligence.

What happened? Better benefits? Stronger network? Broker activity? Acquisition? Pricing? A growing SNP population?

The data does not answer all of those questions. It tells us where to look.

Same with Medigap. Which carriers have gained insured lives? Which have lost them? How concentrated is the state? How did the market look ten years ago?

Or P&C. Where are rates increasing? Where are nonrenewals being reported? Which carriers are tightening? Which ones appear to be opening appetite?

Again, none of this tells an agent exactly what to sell. It should not. If a dashboard tells an agent exactly what to do every time, eventually you probably do not need the agent.

The value is in giving the professional better information and allowing experience and judgment to work from there.

What Will An Insurance Agent Be Paid To Know?

This may be the question that everything else comes back to.

Twenty years ago, access had value. The agent had the carrier, the rate book, the underwriting manual, and the agent knew what products existed. The consumer did not.

That advantage is disappearing. The information is online. Anyone can read it, including the consumer. So simply knowing facts becomes less valuable. However, knowing what they mean becomes more valuable.

There is a difference between 'Carrier A sells this product' and 'Carrier A is becoming aggressive in this market, but based on this client's situation Carrier B is still the better fit.'

Likewise, there is a difference between 'This is the cheapest Plan G' and 'This company is $8 cheaper today, but there are reasons I prefer the carrier B for the long term.'

The same concept applies across all lines of insurance, and the idiosyncrasies of each individual product and carrier.

That is judgment.

And the bad part? AI is going to get better at judgment too. I have no interest in pretending otherwise. But today, and probably for quite a while, the combination of strong technology and experienced professional judgment looks more powerful than either one by itself.

So, Is The Independent Agent Dying?

The evidence does not support that conclusion.

Independent agencies still place about 62% of U.S. P&C premium. We dominate commercial insurance. Their personal-lines share has increased. The strongest independent agencies remain profitable. Investors continue buying insurance distribution. BLS still projects tens of thousands of agent openings every year. Medicare remains complicated. The homeowners market has become harder to place. [1][3][4][5][6][7]

None of that resembles an occupation at the edge of immediate extinction.

But I also do not think the traditional insurance job survives intact.

The person whose value comes primarily from getting a quote, reading product information, entering an application or relaying information that exists somewhere else should be concerned. Technology is already attacking those functions.

The person further up the value chain is in a different position: understanding underwriting, understanding the carrier, understanding the local market, recognizing when something changed, helping a client make a complicated decision, explaining why, and advocating when something goes wrong.

Those are harder problems. They are not impossible for technology. They are simply harder.

The Insurance Intelligence Professional

I have used this phrase throughout the research: insurance intelligence professional.

Maybe that name sticks, maybe it doesn't, I don’t know. The label is not important, but the idea is.

The traditional insurance salesman begins with access: 'I represent these carriers.' The next generation will start somewhere else: 'I understand this market.'

The professional knows where rates are moving, what carriers want, how underwriting differs, where market share is changing, where products are becoming more competitive, where they are becoming less competitive, and what those changes mean for the client.

Technology becomes part of the work.

A Final Expense producer shouldn’t waste twenty minutes searching through underwriting PDFs if software can do it in seconds. A Medicare agent should not be guessing which carrier has been gaining market share when the data can answer the question. A P&C agent shouldn’t discover a carrier appetite change after receiving three declines if the information can be shared in real time. And a medigap agent should not evaluate a carrier solely on today's premium when historical information is available.

None of those tools replaces judgment. They make better judgment possible.

And that may be where the opportunity is.

Where I Think This Goes From Here

The insurance agency of 2036 probably looks different from the insurance agency of 2026.

There may be fewer people doing routine service work. There will probably be more automation. More AI. More structured underwriting data. More market intelligence. More consolidation. Consumers will expect much better digital service. And agents will probably have to justify their place in the transaction more than they have in the past.

I think that is healthy.

If an agent is adding real value, it should be visible. If the customer can replace the agent with a website and lose nothing, then the industry needs to be willing to admit that.

But many insurance decisions do not work that way. There is too much variation. Too many carriers, underwriting appetites, local differences, and that can go wrong. In short, things don’t always fit together neatly.

The independent-agent model survives as long as the independent agent helps solve those problems better than the alternatives. That is the standard. Not history or tradition. Not 'we have always done it this way.' Actual value.

The most important battle over the next decade in the insurance industry may not be between AI and insurance brokers. I suspect it’s going to be small agencies that find themselves competing against independent brokers and large agencies that suddenly operate at scale where they can’t keep up. Certain agencies will be faster, know more than their competitors, and need fewer hands to get the job done.

The agent who uses the tools can research faster. Compare and contrast better. Understand markets and spend less time on paperwork and more time with clients. One agent may eventually be able to handle the work that took several people a generation earlier.

That will eliminate some work. It will also make good professionals much more capable.

I don't think every independent agent wins that transition. Some will sell their agencies. Some will leave. Some will get squeezed by bigger firms. Some products will continue moving toward direct distribution. Some jobs inside agencies will become unnecessary. That is probably unavoidable.

But the evidence does not tell me the independent agent is going away. It tells me the definition of a valuable independent agent is getting narrower. And better.

The future belongs less to the person who simply has information and more to the person who knows what to do with it.

For the independent insurance industry, that may be the biggest change of all.

What The Data Does Not Tell Us

A report like this can become misleading very quickly if every statistic is stretched into a conclusion it does not support.

Independent agencies placing 62% of P&C premium doesn’t mean 62% of consumers prefer independent agents. Premium market share and customer preference are different things. Plenty of consumers still buy insurance online. The increase in independent personal-lines share does not prove that technology failed. Part of the increase may have resulted from market conditions that pushed consumers to shop more aggressively.

The Bureau of Labor Statistics projecting 43,100 openings per year does not mean 43,100 new insurance-agent jobs will be created annually. Many of those openings are replacing older, retiring agents. An aging industry alone will be enough to force a transformation of how agencies operate. Half of insurance brokers are expected to retire between 2021 and 2036, while only one new agent is entering the profession for every six who leave. If agencies want to survive, they will be forced to rethink their business strategy.

That does not necessarily mean agencies should turn to robots. AI users reporting four hours of weekly time savings does not mean every agency will automatically become more profitable.

Medicare Advantage reaching 55% penetration does not mean every county is heavily penetrated. There are huge geographic differences. It is hard to compare Miami-Dade County to the state of Wyoming. And, SNPs producing most of the recent MA growth does not mean the same thing will happen every year going forward.

There is another limitation worth remembering. Premium is not policy count. Policy count is not household count. Enrollment is not profitability. Market share is not quality. Growth is not always healthy. A low rate is not always value. A high rate increase is not automatically proof of poor management.

Data helps answer questions. It does not eliminate the need to ask good ones.

Final Finding

There is one conclusion I am comfortable putting more strongly than the others.

The independent insurance agent is not disappearing. The easy parts of the independent insurance agent's job are.

That difference matters. Because if we confuse the two, we will spend the next decade debating whether technology is going to eliminate agents while the better agents quietly use that same technology to become more capable.

In the past, simply having the carrier appointment and knowing how or where to find information gave agents and brokers the advantage. That’s gone. The advantage now is:

Knowing what to do with the information.

And that is where the next generation of independent insurance agents will either prove their value - or fail to.

---

For references please download the attached PDF