In 2023, Hippo faced what CFO Guy Zeltser describes as perhaps the worst quarter in the insurance company’s history. The response was not to chase growth harder. Instead, leadership made a decision that would have seemed counterintuitive during the company’s earlier expansion: shrink part of the business.
Between 2023 and 2024, homeowners insurance premiums declined approximately 30% as Hippo recalibrated the risks it was willing to carry. For Zeltser, the decision marked a turning point in a broader transformation—one that would move the company toward profitability and a more diversified business model.
Today, Hippo supports more than 60 insurance programs, generating fees and commissions alongside underwriting and investment income. That mix matters. Fee revenue offers greater predictability, while improved underwriting performance gives the company confidence to retain additional insurance risk.
Read More
Zeltser’s approach reflects an unusual career path. Before McKinsey and corporate finance, he worked in his family’s toy business, where customer complaints and everyday transactions made the consequences of business decisions tangible.
At McKinsey, another lesson emerged: even a board-approved strategy could stall when short-term financial targets overwhelmed longer-term investment.
Both experiences inform his current agenda.
Hippo is also deploying AI across claims and finance operations. Zeltser reports that AI now handles more than 80% of initial loss notifications, while finance has shortened its quarterly reporting timeline.
With Hippo targeting a threefold or fourfold increase in profitability between 2026 and 2028, Zeltser faces a familiar tension: delivering quarterly results without sacrificing the investments required to build a larger business.
Made Possible By
CFOTL: For someone new to Hippo, what is the business today? Who does it serve, and where is the opportunity?
Zeltser: Hippo started around ten years ago here in the Bay Area. The original vision was to simplify and improve the shopping experience of people who need home insurance. Through technology, you just click a button. You could get a quote after a minute and complete the shopping experience after three to four minutes.
We have evolved ever since. We’re still in the homeowners business, but we have evolved to be a diversified, technology-native carrier. We have different lines of insurance, personal and commercial, because diversification is the name of the game in insurance. You’re getting much more predictable returns.
Read More
CFOTL: Hippo has chosen to retain more insurance risk. What gave you confidence to make that shift?
Zeltser: We started not participating at all in the risk. But slowly, as we improved our underwriting results and started generating underwriting profit, we said, “Okay, now we can actually enjoy and bear the fruits of our own hard work.”
That gave us the confidence. It’s much more predictable now. As a result, we increase the risk so we can generate even more returns for our shareholders.
CFOTL: Before the financial results arrive, which operating signals tell you whether Hippo is performing as expected?
Zeltser: We started with only one program in the homeowner space, but now we have more than 60. We want to see this number keep growing. And we want to see the overall volume of money flowing through this platform growing.
On the underwriting side, we’re looking at pricing trends. We’re trying to understand whether pricing is going up or down. There’s the pricing side, but there’s also the losses side. I want to see pricing behaving stronger than the losses.
CFOTL: Where are you seeing meaningful business value from AI, rather than just faster work?
Zeltser: More than 80% of what we’re calling first notice of loss—when the customer is calling and needs our help—are being handled today by AI.
Our adjusters were already 30% more efficient six months ago. Now it’s probably 50%. That means we can continue to grow the business. That’s real impact. It’s not just making things faster.
CFOTL: Has AI changed how you evaluate risk, or is human judgment still doing most of the important work?
Zeltser: By and large, human judgment is still doing it, especially when we think about risk management. Unlike other companies that think about risks in terms of downside protection, we’re getting paid for taking risks.
AI is still enabling, it helps all of us, but the decision-making itself still needs humans. Absolutely. And I don’t see that going anywhere.