Gauthaman Krishnamurthy (pictured), chief information officer at Slide Insurance, has a pointed view of how the US insurance industry approaches technology: too many treat transformation as a side project, measure it poorly, and hand it to a single team. At Slide, a technology-enabled property insurer founded in 2021 that has grown into one of the largest coastal carriers in the country, Krishnamurthy is building the counter-argument, one measurable business outcome at a time.
With more than 20 years of technology leadership in insurance, he came to Slide with a specific mandate: embed technology directly into the operations that drive underwriting performance and claims outcomes, not around them.
“Innovation is anything that changes how we think about business decisions and operations on a day-to-day basis,” he said. “But what’s more important is the way in which it’s measured and how those innovative efforts are repeatable.”
Krishnamurthy’s focus is on the core of the insurance value chain. In underwriting, his team is embedding data and analytics directly into workflows, not as a layer on top, but wired into the decision itself. The goal is faster risk selection and tighter underwriting outcomes. In claims, the levers are digital first notice of loss, automation, and data-driven triage to cut cycle time.
On the customer-facing side, Slide is expanding self-service capabilities for policyholders and agents. The framing is practical: reduce friction, give users visibility into their transactions, and stop making people call.
None of this, he argues, belongs to a single leader or team. Technology, architecture, data, and security leads at Slide meet regularly to set the platforms and standards that allow the business to move without losing control. Clear ownership and visible accountability, he says, are what keep momentum from stalling.
“We constantly review progress, make decisions quickly, pivot when we need to, and immediately incorporate whatever we have learned into the roadmap,” he said.
The most candid part of Krishnamurthy’s assessment concerns a failure pattern he has seen throughout his career in the industry: technology initiatives that run well as experiments and die on the way to scale.
“The biggest challenge in integrating any new technology is moving from small pilots or experimentation to operational adoption,” he said. “And especially in an environment like insurance, which is risk-driven and regulated, that has to be at the forefront of everything.”
His response is to treat any technology integration as an operating model change, not a technology project. Business stakeholders are brought in during the design phase. New capabilities are built around existing workflows rather than added on top. The objective is to remove the perception that a new tool is another workload.
The same discipline applies to third-party vendors. Slide uses external partners selectively and requires them to understand insurance operations, not just the software. Krishnamurthy looks for providers with a track record of moving from pilot to full-scale deployment in a regulated environment and who will embed their teams within Slide’s operating model rather than hand off a product.
“We hold our partners to the same standards as any internal project or initiative, clear metrics, defined ownership, transparent decision making, frequent conversations to align direction, and a clear path from concept to production,” he said.
The question of whether insurers should build proprietary large language models, so-called agentic AI systems capable of autonomous decision-making, or buy commercial products is generating significant debate across the industry in 2026. Krishnamurthy is watching carefully and is not ready to declare a winner.
“The space is evolving rapidly, there are many vendors available, and it’s important to understand what each one brings to the table,” he said. “But as insurance carriers, we have to be responsible about what we are adopting and how.”
He pushes back on the binary framing of build versus buy. A third model, he suggests, may already be forming, one in which core insurance platforms begin to deliver AI capabilities natively, sitting alongside both proprietary builds and commercial tools. The implications for carriers’ technology stacks, and for the vendors competing for those budgets, are significant.
“It’s still a bit early to fully determine whether this is the old model of build versus buy, or whether a third model is emerging, one where there’s a coexistence of the two, where core platforms may also begin providing some of those AI capabilities natively.”
That caution is not inaction. It is a deliberate posture in an industry where a badly implemented model, one that influences underwriting decisions or claims outcomes, carries regulatory and reputational consequences that a software rollback cannot fix.
The through-line in Krishnamurthy’s approach is consistency: whether the conversation is about agentic AI, core systems, or self-service tools, the answer starts with the business problem, not the technology.
