Zywave’s new Broker of the Future study says AI is set to unlock more than $100 billion a year in value for carriers and brokers over the next five to seven years — and it says the broker is not the thing being replaced. In Zywave’s own words, no insurer has emerged to underwrite an AI system’s mistake at the point of bind. The licensed human stays in the review seat. The software takes the paperwork.
That is the whole shape of the change in one sentence: AI does the volume work, and a licensed person still owns the decisions.
The gap is in the economics, not the headcount
Zywave’s starkest number is revenue per employee. A traditional insurance agency, the study says, runs on roughly $175,000 of revenue per employee. An AI-native brokerage is proving capable of running at closer to $1,000,000 — nearly six times as much.
Read it the way Zywave frames it: these firms are growing their business faster than their headcount. That is not a layoff story. It is a throughput story. The firms pulling away are the ones whose people spend their hours on clients and judgment instead of re-keying the same information into a fourth system.
Zywave is a technology company that sells into this market, so read its study the way you would read any vendor’s research — the direction is worth more than the decimal points. But the direction matches what anyone who has renewed a policy already knows: the paperwork load in insurance distribution is enormous, and most of it is not judgment.
Where Zywave says the $100 billion sits
The study breaks the opportunity into four shifts, all of them already underway by its account:
- Pricing on live data — $10 to $20 billion a year, in Zywave’s estimate. Continuous underwriting against real-time inputs instead of a once-a-year snapshot of your business.
- Software taking the back office — $30 to $40 billion a year. The clerical core of distribution work, which Zywave says occupies more than half the industry’s workforce today.
- Better risk selection and fraud detection — $20 to $30 billion a year. Zywave notes carriers are already paying significantly more in contingent commission for a well-selected book — which means a broker who brings clean, well-chosen risks gets paid for it.
- New AI risks becoming new coverage — $15 to $20 billion a year. A new specialty line that Zywave compares to cyber insurance’s run from zero to $15 billion in premium.
Notice what is not on that list: a shift where the software absorbs the consequences of being wrong. Every dollar above is volume, speed, selection, or a new product. None of it is accountability.
The line the software doesn’t cross
Zywave puts the broker’s real value plainly: it is not processing forms. It is judgment — “uncovering the hidden context between what a client says they need and what they’re actually exposed to.” Insurance, the study says, remains a high-consequence product that must be sold, not bought.
The point of bind is where that stops being philosophy. Binding is the moment coverage becomes a promise. If the exposure was described wrong, if a classification was off, if the thing the client forgot to mention turns out to be the thing that burns — somebody answers for that. Zywave’s finding is that no insurer has stepped forward to be that somebody for an AI system’s error. Until one does, the economics can change completely and the accountability stays exactly where it is: with the licensed human who reviewed it.
That is not an argument against the software. It is the study’s own argument for using it correctly. As Zywave’s CEO Martin Simoncic puts it, the firms operating at a different level are the ones that “put AI to work on the paperwork rather than the judgment.”
The shortage is the other half of the story
The study’s real warning is not replacement. It is retirement. Zywave, citing Bureau of Labor Statistics projections, says roughly 400,000 insurance professionals will have retired since 2021 by the end of this year, with not enough new talent behind them to replace the judgment and experience walking out the door.
Put the two halves together and the picture gets clearer. The industry is losing experienced judgment at the exact moment the economics reward firms that can scale. The brokerages that win, on Zywave’s account, are the ones that use software to stretch the judgment they still have — and the study describes three shapes emerging: the biggest firms consolidating premium and talent, small and mid-sized agencies that don’t modernize becoming acquisition targets, and specialists pulling away by going deeper into a niche. Zywave’s product chief, Eric Rentsch, adds the unglamorous part: the winners “aren’t necessarily the ones spending the most on AI, they’re the ones who invest to get their own data in order first.”
What to ask your broker before the next renewal
You probably do not care which software your broker runs. You care about two things at renewal: whether anyone with a license actually reviewed your account, and whether the submission going to underwriters reflects what your company really does. The hidden context Zywave describes — the gap between what a client says they need and what they’re exposed to — is where renewals are won and lost, and it is exactly the ground covered in our conversation with a producer about where contractors actually have room at renewal.
So ask the plain question: what does the software do at your firm, and what does a licensed person still check by hand before anything binds? A good brokerage will answer that in one breath. That answer — volume on the machines, judgment in the review seat — is the broker of the future Zywave is describing. The study says the economics of that shape are nearly six times stronger. The bind still needs a name on it, and that part has not changed.