AI in Insurance: Building Trust with TPAs and Smarter Claims Processing
BY EDDIE OLSONWith any new technological advancement, the initial understanding can sometimes be difficult to grasp. For example, “AI for enterprise” sounds great, but clearly explaining its use for business might be challenging at first. I remember the exact moment the new and abstract idea of blockchain for business made sense to me, not in theory, but in practice.
A founder was pitching a platform that could instantly verify worker credentials at the moment a contract needed to be triggered. It was fast, simple, and practical. I told him, “If you don’t mention that this runs on blockchain, your customers will never know.” He smiled and said, “Exactly.”
That stuck with me. The best technology doesn’t announce itself. It disappears into the experience, not as a set feature, but as a greater part of the product or process. The same can be true for AI in enterprise, and in this case, insurance.
Navigating Insurtech AI Use Cases
Insurers are leaning hard into AI to scale operations, reduce friction and improve performance across underwriting, claims and customer service. This initiative may come from a transformation mandate, business pressure or competitive fear; and, though their efforts are well intentioned, many with strong use cases, many fall short.
You may have seen the recent Fortune Article explaining how 95% of generative AI pilots at companies are failing. Additionally, according to BCG, only 7% of insurance companies surveyed have successfully brought their AI systems to scale. That’s not because the technology doesn’t work; rather, it’s because most try to automate complexity.
I’ve worked alongside startups, insurers and operators trying to bring meaningful change to the industry. Across the complexity of it all, one thing holds true: AI can’t fix broken promises. It can only scale them.
Making AI Work
1. Clarity That Builds Trust
Insurance is, at its core, a promise, one that people rely on when things go wrong. But too often, that promise gets buried in complexity, dense language, vague exclusions and product designs that sound smart but feel risky. When people don’t understand what they’re buying, they don’t trust it. And when AI enters the mix without clarity, it becomes a liability.
The best AI efforts in insurance don’t add features. They remove friction, this could include: explaining coverage in plain language or conversation; highlighting what’s included, not just what’s excluded; building custom experiences for the policy and policy holder.
Clarity isn’t just good UX, it’s building a foundation of trust from the start, and this trust is what makes transformation stick.
2. Clean, Governed, Owned Data
Alexander Wang, CEO of Scale AI, mentioned recently on a podcast that AI’s three key pillars are data, compute, and algorithms. Insurers are sitting on a gold mine of data and have a massive opportunity to leverage their data with AI.
The early applications of the data in AI are still in their infancy, which can lead to potential pricing errors, claim delays and systemic risk across underwriting. Legacy platforms and policies exacerbate the problems for insurance professionals due to manual inputs, duplicate records, and no version control.
Yet, what separates high-performing teams is that they treat data as infrastructure, which is critical in AI-first foundational platforms. A data-disciplined culture that treats accuracy as a priority will lead to strong AI applications for the organization.
3. A Strategy That Fixes the Foundation First
AI works when it scales something that’s already sound. You can’t expect intelligent automation from unintelligent systems.
That means, to best begin AI implementation, focus on core areas of impact, such as: fixing what’s slow, inconsistent, or manual (that can be automated); standardizing how data flows from quote to claim and applying best data management practices; removing redundancy, removing illogical steps and focusing on accuracy prior to scaling.
As many insurers focus on AI as the transformation, some of the best ones understand it’s the amplifier, not the foundation.
TPA(I)s - Why TPAs Are A Great Place for Enterprise AI
Third-party administrators (TPAs) often handle a very important part of the insurance journey, such as the claim. And in many cases, they do it using outdated systems, manual processes, and fragmented workflows… perhaps the perfect place for AI implementation.
TPAs are becoming one of the most important delivery layers for AI in insurance. Why? Because they touch everything AI needs to be effective: clean data, operational workflow, and customer outcomes.
They’re no longer just administrators. They’re infrastructure. In 2025, we’ve seen a wave of capital flow into AI-native TPAs, each reimagining claims from the ground up:
Reserv raised $41 million in Series B funding, led by QBE Ventures, Bain Capital Ventures and Flourish Ventures. Their model uses AI to streamline claims, automate documentation, and produce real-time insights. They operate as a full-stack TPA for digital-first carriers and MGAs.
Elysian, raised $6 million in August 2025 from Portag3, American Family Ventures and TenOneTen Ventures. They’re truly AI first, using AI to bridge gaps in policy logic, reduce adjustment time and increase transparency for brokers and customers.
These companies are building the infrastructure where customer trust is delivered, claims. Though the modern customer expects a modern (faster) claim experience, it’s also about simplicity, removing ambiguity and restoring confidence.
Conclusion: Simplicity Wins. Strategy Scales.
The future of insurance belongs to those who simplify the experience, own their data, and modernize the infrastructure where trust is tested. Multimodal, (2024 gener8tor alumni) builds and manages secure, integrated, and tailored Gen AI automation for complex workflows in financial services. They shared their vision that I think is a great summary and position for success:
“We founded Multimodal with a clear belief: AI is at its best when it disappears into the way insurance teams already work. Our systems are built to orchestrate entire workflows, not just automate individual tasks, removing noise, cutting out unnecessary steps, and giving both adjusters and policyholders the confidence that every decision is clear, accurate, and fast. For TPAs and insurers alike, that’s how we turn a promise into trust, day after day. “
Today is a good day to focus on data discipline and clear outcomes.
Thanks for reading!
- Eddie Olson
