On February 9, 2026, insurance broker stocks suffered one of their worst single-session declines in years. Willis Towers Watson fell 12% — its worst session since November 2008. Aon hit a 52-week low. Arthur J. Gallagher dropped nearly 10%. The catalyst: two AI-powered insurance distribution tools launched simultaneously inside OpenAI's ChatGPT, putting personalised insurance quotes in front of 800 million weekly users.
Markets saw disruption and sold first. The wiser question is: disruption of what, exactly — and for whom?
-12.0%
Willis Towers Watson
Worst session since 2008
-9.9%
Arthur J. Gallagher
-9.3%
Aon
Hit 52-week low
-7.0%
Marsh McLennan
The critical distinction markets missed
The AI tools that triggered the sell-off — Insurify's comparison app and Tuio's direct quoting capability — both target personal lines: auto and home insurance. This is a commoditised, high-volume, low-complexity market where comparison and price transparency have genuine consumer value.
Here is the critical fact markets appeared to overlook: Marsh McLennan, Aon, and Willis Towers Watson derive between 80% and 90% of their revenues from commercial lines, specialty insurance, reinsurance, and consulting. These are not businesses that a ChatGPT comparison widget can disintermediate.
"The sell-off was overblown. ChatGPT apps target personal lines while the major public brokers derive most of their revenue from commercial lines requiring human expertise."
— Wolfe Research, February 2026
Where the revenue actually comes from
To understand why the sell-off was an overreaction, you need to look at the revenue structure of the three largest brokers.
Marsh McLennan
$26.2bn
51% commercial broking, 10% reinsurance, 37% consulting
Aon
$17.2bn
54% commercial risk, 13% reinsurance, 33% human capital
Willis Towers Watson
$9.7bn
48% corporate risk & broking, 32% health & career
Willis Towers Watson's story is particularly instructive. The firm sold its only significant direct-to-consumer business — TRANZACT, a Medicare insurance distribution platform — in January 2025, realising a pre-tax impairment loss of over $1 billion. That exit, painful at the time, now looks prescient. WTW's leadership had already concluded that personal lines distribution was a non-core, structurally challenged business.
AI disruption risk by segment
Not all broker revenue is equally exposed. The risk varies significantly by segment.
Segment
AI disruption risk
Large commercial risk broking (Fortune 500)
Very LowSpecialty / E&S — cyber, D&O, construction
Very LowReinsurance broking
Very LowMiddle-market commercial P&C
Low–MediumSME / small commercial retail broking
MediumEmployee benefits administration
MediumDirect-to-consumer personal lines
HighThe real AI story: augmentation, not displacement
For specialty insurers, reinsurers, and large commercial brokers, the more relevant AI story is entirely internal. McKinsey estimates generative AI could unlock $50–70 billion of insurance industry revenue globally, with the highest near-term impact in marketing, sales support, and customer operations.
Early evidence is compelling. One US agency deploying an AI platform reported average staff time savings of 7.5 hours per week and a 244% ROI. Aon has partnered with DataRobot on agentic AI for client onboarding and servicing. Marsh McLennan launched its internal generative AI tool, LenAI. WTW is investing in AI across its Corporate Risk & Broking and Insurance Consulting & Technology segments.
The investment thesis for large brokers is therefore AI as operating leverage — revenue per producer rising as service costs fall — rather than a story of displacement.
Three questions every board should be asking
Where is our personal lines exposure?
Not all brokers are equal. Firms with meaningful SME or retail books face genuine near-term pressure on commission economics. Quantify it honestly before the market does it for you.
Are we capturing AI as operating leverage?
The efficiency gains available through AI in underwriting support, claims, and client servicing are substantial. The question is not whether to adopt — it is how fast and where first.
What new risks are we positioned to underwrite?
AI liability, autonomous systems, and algorithmic risk are emerging as significant new specialty lines. Early movers in these categories will define the next generation of specialty insurance.
Summary
The February 9 sell-off was a market overreaction to a personal lines distribution innovation. MMC, Aon, and WTW derive 80–90% of revenues from commercial lines, specialty, reinsurance, and consulting — all with low near-term disintermediation risk. For large brokers, AI is overwhelmingly a story of augmentation and operating leverage, not displacement. The genuinely exposed segments — SME retail and personal lines — represent a small fraction of the major brokers' revenue base.
Sources: Marsh McLennan Q4 2024 earnings; Aon 2024 10-K; WTW Q4 2024 earnings; MarshBerry Broker Composite Index; McKinsey AI in Insurance (Feb 2026); Wolfe Research; Bloomberg Intelligence. This article does not constitute investment advice.