Insurance is navigating a generational handover just as artificial intelligence adoption accelerates, and Tatum Fish (pictured), strategic value architect at Cytora, says the industry's most effective AI deployments have nothing to do with replacing underwriters and everything to do with clearing the administrative work in front of them.
Having worked inside a global carrier before moving into insurtech, Fish said the goal should never be automating the underwriter but automating the tasks that keep the underwriter from doing what only they can do.
"It's not about saying we need AI everywhere," Fish said. "It's more about asking: where are people spending time today that doesn't require their expertise, so that we can bring AI into the process to give them more time and efficiency."
That distinction matters across specialty and commercial lines, where highly skilled staff often spend hours moving information between systems, reading documents and rekeying data rather than applying the expertise they were hired for. Fish pointed to Applied Systems' email to quote tool as an example of automation aimed squarely at that administrative layer, aggregating submissions that arrive through email, phone calls, texts and carrier portals so underwriters can review risk rather than key it in.
"The insurers and brokers that are going to be most successful with AI are the ones connecting technology to a realized business problem," Fish said, adding that involving the people doing the work is central to building solutions that actually get used, a theme explored further in Insurance Business America's recent report on the industry's underwriting talent shortage.
Fish said the most persistent misconception in the carrier-insurtech relationship is the assumption that the other side does not understand the problem. Technology companies can view carriers as resistant to change, while insurers can dismiss a strong demo as unworkable in production.
“The goal is not to automate the underwriter – the goal is to automate the work that prevents the underwriter from doing what only they can do.”
"Insurers aren't resistant to change – there are processes built over decades that have created genuine barriers to acting swiftly," Fish said. At the same time, Fish said insurers often assume new tools will work out of the box without co-design, when technology only performs as intended once it is guided toward the specific process a carrier is hoping to achieve, a tension underwriters continue to navigate as automation expands, according to Insurance Business America's coverage of AI trust levels in underwriting.
Asked how insurers can adopt automation without hollowing out the judgment built over decades, Fish returned to the same principle.
"The goal is not to automate the underwriter – the goal is to automate the work that prevents the underwriter from doing what only they can do," Fish said. "There's broad alignment in the industry that nobody wants AI to be a black-box decision maker. Nobody wants it to replace years of underwriting experience."
Fish said mid-career staff running daily workflows are rarely the ones shaping technology strategy, even though they hold the clearest view of where automation will help or hinder. "In my experience, 99% of the time, the person who can tell you where technology is going to help or create a hindrance is the person doing the task manually right now," Fish said, a pattern echoed in Insurance Business America's reporting on how AI is reshaping day-to-day insurance roles.
That involvement, Fish added, changes the narrative from insurers dictating how AI will be used to building tools with employees in mind from the start.
For brokers and underwriters uneasy about AI reshaping their roles after a decade or more in the industry, Fish's advice is direct.
"Don't underestimate your knowledge," Fish said. "AI can process information very quickly, and it's changing at a rapid pace, but it doesn't replace judgment, relationships, or the contacts you've built over 10 or 15 years in the industry."
"Get comfortable and get curious about AI rather than trying to avoid it," Fish said. "You should understand what the tools can do and start playing with them to see how they can make you better at your job."
