What’s AI’s role in brokerage M&A?

While insurance brokerages seek ways of using artificial intelligence (AI) to enhance their efficiency, firms engaged in brokerage mergers and acquisitions (M&A) are also harnessing the technology to build deals.

Right now, a lot of that AI use is indirect, says Andrew Mathias, managing director and partner of KPMG Corporate Finance, the firm’s investment banking group.

It’s used via M&A subscription platforms, such as Capital IQ, PitchBook and others, that use AI on their back ends. The platforms scrape the web for company data and gather other intelligence.

“We use those platforms religiously, probably every day in some way, shape or form,” he tells CU.

“We use AI to scrape the market and give us intel, whether it’s deal size or just transactions that are happening. We also use it for targets. I do a bunch of work on the buy side for larger players, whether they’re insurance companies or brokerages or global MGAs. We use AI to help us supplement…when we’re looking for targets or to help build our landscape.”

It’s also useful in Mathias’ sell-side work for building and updating a list of 2,500 targets he originally developed in 2020. Back then, he hired a team of 10 summer interns and had them spend two-plus months scouring the web for details on companies that could be open to mergers.

“Now I don’t have to do that because…we have a full deal team that just focuses on coding and utilizing AI to do things like this. And I had them help me build a program to scrape the market and do it a little bit more efficiently,” he says.

“It’s not perfect, and I have to review it and then go back and help refine it. But that’s kind of the way that we’ve utilized [AI] on the M&A side.”

Scraping the market

Unlike European and to a lesser extent U.S. markets, where it’s easier to come by private information about companies, Canada’s markets can be opaque.

“Part of the Canadian market is extremely private, especially on the brokerage side or the distribution side, and wholesale intermediaries as well,” Mathias tells CU.

Using AI to find data on revenues, direct premium written (DPW) and gross written premium (GWP) can help M&A strategists determine potential merger candidates. But there can be a catch.

“The AI is good, but it’s logic driven. There’s no judgment, and so it’ll come back with revenue but then when you dig deeper, that revenue’s from 2016 and…is just not relevant to 2025,” he says.

“What I use as a primary driver, especially in today’s market, is number of people. It’s usually a good indicator – producers, brokers, etc. And then market access is [a data point] that’s very important in today’s world. If they’ve got two general markets and then they’ve got 20 MGAs, they’re probably not of a specific size [for an M&A target].”

Currently, Mathias says, some larger carriers are less willing to work with smaller brokerages and in certain cases are dropping them. So an acquisition target’s market count is a useful data point when M&A consultants can’t obtain its earnings before income, taxes, depreciation and amortization (EBITDA), or premium data.

“When you’re doing that search, you can sometimes search for, ‘Does this brokerage have this market?’ If they don’t, then it falls under a certain size threshold,” he says.

Feature image by iStock/Yuuji