How agentic AI is rewiring insurance for 2026

Despite the reluctance of many insurance companies to fully adopt artificial intelligence, the expert consensus is that agentic AI is poised to take center stage in the insurance industry in 2026.

Inci Kaya, research manager, IDC Financial Insights, suggested a growing number of insurers will make slow, steady progress to get there over the course of the next year. 

“Agentic AI and Gen AI, all flavors, are very top of mind,” Kaya said. “What I’m seeing is that carriers know what they want. They are not getting taken away by all the shine and gloss of generative AI and everything AI. I think you’re going to see a steady and clear-eyed, deliberate, perhaps on the slower side, but deliberate approach as to how they want to go about it.”

Agentic AI is the next step beyond today’s chatbots — it doesn’t just answer questions, it can take action on its own. In other words, it can plan, make decisions and carry out tasks autonomously, all while staying under human supervision.

In an interview with Insurance News Net, Kaya noted that many carriers have launched pilot AI programs, testing out specific use cases rather than opting for enterprisewide solutions. While some experts have argued against this approach, Kaya believes many companies will continue to refrain from having “everything full-blown AI,” instead opting for “predictive analytics plus a sprinkle of AI.”

Nonetheless, experts agree that insurers will continue to invest in AI, even in the face of recession fears and particularly as consumers, who themselves use tools like ChatGPT and Claude on a daily basis, increasingly trust it.

“I don’t think the insurance companies are going to take a backseat. Initially, they’ll be risk averse, like financial services companies always are, but I think this is one of those times where it’s just too obvious and too powerful and too intuitive to sit back,” Alex Sion, head of financial services, Blend360, added.

From AI co-pilots to AI agents

Franklin Manchester, principal global insurance advisor, SAS, told Insurance News Net that insurers could begin phasing out policy admin systems in favor of Gen AI “assistants” built on large language models such as ChatGPT and Claude by next year.

Many carriers have already developed and engaged AI insurance co-pilots such as this, establishing a distinct market for AI-powered tools specific to the insurance industry.

Manchester explained that policy admin systems require a substantial amount of investment and upkeep, “but you don’t actually need them to interact with your data if you’re running a co-pilot.” Rather, the co-pilot can connect directly to data just like the PAS does.

“Sooner or later, a CEO and tech team are going to figure that out — and when they do, they’re not going to need the admin system in order to do their day-to-day job,” Manchester said.

And Sion believes things will go even further in 2026, with tech-forward carriers graduating from AI assistants to full-blown agentic AI.

“The next horizon we’re moving into is the world of agentic AI, where AI can now not only do the Gen-AI-oriented kind of question-and-answer and smart research thing, but it can also now drive and execute tasks and organize itself autonomously to not only suggest tasks but then have the power and the authority and the integration to execute them,” Sion said. 

He said that while generative AI “turbocharged the productivity of humans,” agentic AI will be capable of transforming organizational processes and workflows in a “much more fundamental way.”

“Now you’re either skipping human steps or eliminating them altogether and replacing them with a different kind of process and a workflow that is run more by agents — still managed always by humans, but run more autonomously by agents,” Sion said.

He described agentic AI as a “massive unlock” for insurance as “one of the most process-regulated, paper-heavy industries.”

“The idea of potentially transforming the way insurance is distributed using AI-enabled conversational interfaces is not only intuitive, but it’s massive. You don’t have to think that hard to imagine a future where agents are either dramatically augmented by AI assistance or even replaced, and that technology is moving very quickly,” he said. 

Is insurance ready for agentic AI?

However, Manchester believes “not many insurers here in the U.S. or globally are ready for agentic AI” just yet — and won’t be until late 2026.

“Insurers are not really mature yet on their agentic AI journey. I think they’ll start seeing use cases mature in the marketplace by the end of 2026, but it will be on what I might call low-downside-risk use cases,” he said.

Manchester doesn’t believe insurance will get to the point where AI agents make complete determinations without human oversight, such as when handling claims — at least for the foreseeable future. Rather, he said insurers looking to adopt agentic AI should focus on aligning digital agents with steps and processes that come alongside humans with the goal of streamlining those processes.

“You have likely heard the term ‘humans in the loop.’ I want to flip that idea on its head. I want to talk about AI in the loop to figure out when you’re using agentic AI to do certain things and very long insurance processes that don’t create downside risk for you,” he said.

He likened this to agentic AI functioning like a claims processor or an “entry-level position at an insurance company, where you have someone that’s helping facilitate the overall claims process, but the most experienced person in that process flow is the human and they have supervisor authority over a group of agents who are doing that process.”

‘Housekeeping’ needed: upskilling and data cleaning

Kaya pointed out that adopting AI could have a spinoff impact on the workforce in 2026, which follows a thread similar to what Manchester pointed out — namely that the “entry-level” role being assumed by agentic AI could deprive newcomers of that hands-on experience.

As such, she projects that workforce upskilling will accompany increased adoption of AI over the next year and well into the future, even as the industry continues to grapple with a yearslong talent gap crisis.

“Number one, there aren’t enough people that are interested in joining. Number two, the tools are not necessarily in place to make it appealing for the younger generation to join the ranks. And number three, because of AI, we’re also losing that entry-level, hands-on practice chance for incoming employees to learn the trade,” Kaya said.

She said that while it is nice to have AI automate low-grade processes, “the downside or the underbelly of that is that the new people coming in are kind of being deprived of the opportunity to learn hands-on.”

“Now, are there AI tools like ChatGPT and Copilot and this and that to help them with that? Yes, there are. But the big question mark — and I’m not convinced of this — is are those insights that the underwriters have developed over the years available in a co-pilot setting for the new people to learn automatically? I’m not convinced of that. I’d like to see some more proof of that,” she said.

Taking a step back even further, she also noted that many businesses “need to get some housekeeping done” to even prepare those AI systems to function effectively, such as structuring data and establishing governance structures.

“If you can do that, you’re going to get a lot of mileage out of AI. … Governance and data are the two key things that I would like to see carriers lock down,” Kaya said. “I think that’s not going to be a one-year effort, but it’s definitely step one. Without those things, we’re just talking; we’re just playing pretend.”

Sustained investment

Performing the recommended “housekeeping” to prepare to adopt AI requires time and investments, but Kaya said most carriers are not looking to put an end to that. Interestingly, carriers are sustaining AI investments even in the face of economic concerns.

According to an IDC survey, 54% of American insurance carriers believe an economic recession is likely within the next 12 months. Even further, they’re “looking at geopolitical conflicts, economic and political uncertainty, tariffs and cyberthreats to add to the mix.” 

Despite this, IDC’s research found that survey respondents did not plan to reduce their budget allocations for AI-related projects.

“If they need to adjust something, they are adjusting their business revenue expectations downward if need be, but they’re not pulling money out of AI. They might be cutting out some other categories of spending, but the AI-related spending is not likely to be affected, just based on what we’ve been seeing,” Kaya said.

She suggested this could be because their top priority is to grow their business and their second top priority is expanding their customer base, both of which “give us a hint as to what the AI-related use cases might be.”

Changing consumer behavior trends

Sion and Manchester similarly believe carriers will leverage AI to expand their customer base in 2026, largely because clients themselves are increasingly using AI tools to research and shop for insurance products.

“The consumer side on this front, and changing consumers’ preferences in terms of how they shop and explore the world, is going to be the catalyst that moves things faster within the industry,” Sion said.

He noted that “the idea of AI-assisted research” is already gaining popularity as more consumers are using AI tools like ChatGPT or Gemini for research instead of searching Google and clicking on web pages. For instance, a customer could use a chatbot to “ask 500 questions about the most detailed nuances of insurance.”

For his part, Manchester believes one in every two Americans will use this “zero-click research” method to shop for insurance online going forward.

“My prediction is that, in 2026, half of all U.S. insurance consumers are going to use AI tools to research and shop for insurance policies. And SAS’ own data from the trust imperative survey we just did with IDC actually shows that survey respondents — insurers, decision-makers — trust generative AI 100% more than machine learning,” Manchester said.

“It’s extremely powerful, and the interfaces will force insurance companies to think differently about the way that they deal with sales, marketing and distribution on the digital channel alone,” Sion added.

However, Manchester also noted a potential downside to AI-savvy customers: bad actors who can use it for fraudulent ends. He believes the industry will see an increase in AI-related fraud in 2026 as usage, and expertise, ramps up.

“We are starting to see alarming reports about the number of claims that have been submitted using AI forgeries, so using generative AI to create documents, text, video, emails, et cetera,” he said.

He cited data from the National Insurance Crime Bureau that estimates around 10% of insurance claims contain some element of fraud, and said he believes that figure will double over the next year.

“I expect that number, in 2026, is going to double. Insurers will see 20% of all claims contain some element of fraud due to generative AI. These tools are becoming ubiquitous, and they’re showing up more and more from a fraud perspective, not just here in the U.S., but abroad in other countries where insurers are doing business,” Manchester said.

An AI-powered future

In the face of opportunities and challenges, Kaya is optimistic about where the insurance industry will take AI in 2026. She noted that while carriers have had a slow, cautious start, many have gone from feeling skeptical to feeling more confident. 

“If you look at the conferences and industries and all the speaking topics, the issues have evolved away from bias and hallucination and they have evolved more towards how can we capitalize on this more solidly,” she said.

She acknowledged that successful AI adoption will depend on leaders getting buy-in from their teams and cultivating “supportive sentiment in the company to make sure people aren’t feeling like they’re being left behind or their jobs are at stake.”

“I think the readiness is there,” Kaya said. “It’s just a matter of internally identifying the use cases they want, getting their data house in order and governance in order, maybe getting some kind of a chief AI officer who’s going to be able to manage and understand both the business considerations and the technical considerations. If you have those, I think you’re in good shape to reap the benefits.”

IDC Financial Insights is a global research and advisory firm that provides data analysis and consulting services to financial institutions. It was founded in 1964 and is based out of Massachusetts. 

Blend360 is a global services company that provides AI solutions and transformation for financial institutions and organizations. It was founded in 2015 and is based out of Maryland.

SAS is an AI, data management and analytics organization founded in 1966 at North Carolina State University.

Rayne Morgan is a journalist, copywriter, and editor with over 10 years' combined experience in digital content and print media. You can reach her at [email protected].