Intact Puts AI Pricing Earnings at C$250M and Says It Will Double

Intact Financial said on Wednesday, November 4, that artificial intelligence earns it close to C$250 million a year in recurring pre-tax earnings and that it expects to double that within 24 months, on a third-quarter call in which it reported a 94.0% combined ratio against 89.8% a year earlier. Machine learning now prices 94% of its Canadian personal lines business and 75% of its commercial lines, chief executive Charles Brindamour said.

The figures came a day after the Toronto carrier released results showing net operating income per share of C$3.63 against C$4.46 in the third quarter of 2025, net income down 35% to C$562 million, and a trailing 12-month operating ROE of 15.8% against 19.6%. Catastrophe losses were C$660 million in the quarter, four points above expectations, chief financial officer Ken Anderson said on the call.

The disclosure comes five months after this site reported that Intact was one of three insurers out of 30 in the Evident AI Index to put a financial figure on AI, and that none of the three had said whether the figure sat in the loss ratio or the expense ratio. Wednesday's call answered that question for Intact.

Key Takeaways

Machine learning prices 94% of Intact's Canadian personal lines and 75% of its commercial lines, with more than 600 proprietary models in production.
About 60% of the planned increase to C$500 million will come through the loss ratio and the rest through the expense ratio, COO Patrick Barbeau said.
The quarter's combined ratio was 94.0% against 89.8%, with catastrophe losses of C$660 million running four points above expectations.
Net operating income per share was C$3.63 against C$4.46 a year earlier, with underwriting income down 38% to C$368 million.
Canadian personal auto ran a 91.2% combined ratio on premiums up 10%; personal property ran 105.0% with 33 points of catastrophe losses.

The pricing figures

Brindamour opened the AI section of his prepared remarks with the penetration figures. "Today, machine learning prices 94% of our Canadian personal lines business and 75% of our commercial lines," he said. The company has more than 600 of its own models in production across Canada, the U.S. and the U.K. and Ireland, and a team of over 600 AI experts feeding them, he said.

He tied the models to claims data. "We handle more than 95% of claims in-house, so we see every step of the claims journey from first notice to final repair," he said, adding that this gives the company data no one else has. "Simply put, you can't prompt your way to a sophisticated pricing model without the ingredients that we have at scale."

The earnings figure followed. "We earn today close to CAD 250 million in recurring pre-tax earnings and expect to double that in the next 24 months," Brindamour said. Asked later how the number is built, he said: "The numbers we're talking about are the numbers we measure, control sample versus the rest." The call gave no base period for the C$250 million.

Investor slides summarised by Investing.com listed the operating results behind the figure: broker quote automation in Canadian commercial lines has cut turnaround times by 70%, advanced pricing models cover 50% of global specialty lines premium, and U.K. property models are expected to support about 50% of risk-based pricing by the end of 2026.

Where the next C$250 million lands

Bart Dziarski of RBC Capital Markets asked how much of the AI benefit would fall to the bottom line against being reinvested over the next two years. Barbeau, the chief operating officer, said about 60% of the increase from C$250 million to C$500 million would come through the loss ratio and the other part would help the expense ratio. "Our top priority is still to invest a good portion of AI into the pricing and risk selection area," he said, listing customer experience, software development speed and efficiency behind it. "We definitely think it will expand our combined ratio outperformance over the next two years."

Brindamour added that "the expansion of the outperformance is bigger than what we measure, to be clear."

The quarter those figures sit against produced underwriting income of C$368 million, down 38% from C$598 million, and net operating income of C$641 million against C$797 million, according to the release. The consolidated expense ratio was 33.9%, a small improvement year over year, and the underlying current accident-year loss ratio was 55.4%, Anderson said. Favourable prior-year development contributed six points. Anderson expects the full-year expense ratio at the higher end of the 33% to 34% range.

The catastrophe loss ratio was 10.7% against 6.8% a year earlier, according to the Investing.com slides summary. Operating direct premiums written grew 5% in constant currency to C$7.0 billion.

Segment, Q3 2026	Operating DPW	DPW growth	Combined ratio	Combined ratio, Q3 2025
Canada	C$4,939M	7%	95.1%	89.1%
UK & Ireland	C$1,078M	Flat	94.1%	95.5%
US	C$989M	3%	85.2%	83.6%
Total	C$7,006M	5%	94.0%	89.8%

Sources: Intact Financial Corporation, Q3 2026 results release (November 3, 2026). DPW growth is in constant currency.

Within Canada, personal auto ran a 91.2% combined ratio on premiums up 10%, with two points of unit growth. Personal property ran 105.0% with 33 points of catastrophe losses, on premiums up 7%. Commercial lines ran 91.9% on premium growth of 2%, with catastrophes seven points above expectations. "Excluding this, the 85% combined ratio was right in line with our five-year average," Brindamour said. The U.S. segment's 85.2% compares with 83.6% a year earlier, and the U.K. and Ireland's 94.1% was broadly in line with expectations, the release said.

The agent question

Brian Meredith of UBS asked how AI personal shopping agents would roll out and whether they posed a disintermediation risk to Intact's distribution system. "Agents will come," Brindamour said. "Fundamental, in my view, is the quality of your digital experience and your digital channel."

He gave three reasons the company expects to hold its position. It has built two of the strongest brands in the Canadian marketplace, he said. It has a digital channel, "unlike many broker-distributed business." And "we own a broker that has close to CAD 6 billion of premium in the Canadian marketplace." He expected pressure in parts of the channel where a broker lacks scale, and said the company is "actively focused on ensuring we win in LLMs as well as with agents in time."

Digital sales premium grew 39% year over year, with about one million monthly active users for three consecutive months, according to the Investing.com slides summary. Distribution income rose 16% to C$171 million in the quarter, the release said. The question follows Allstate's disclosure that its own AI agent is closing policies in three states, as this site reported in June.

"Some believe generative AI will level the playing field. We strongly believe the opposite," Brindamour said in his prepared remarks. "We firmly see ourselves as relative and absolute winners in an AI world." The board declared a quarterly dividend of C$1.47 per common share, payable December 31 to holders of record on December 17.

Intact Puts AI Pricing Earnings at C$250M and Says It Will Double

Key Takeaways

The pricing figures

Where the next C$250 million lands

The agent question

Further Reading

Sources