Hello, FinTech community. Welcome to our FinTech Top Voice Interview Series.
The latest FinTech Interview with Tom Rasmussen, VP Product Claims at Carpe Data is an interactive Q&A-styled conversation.
Tom Rasmussen is the Vice President of Product for Claims at Carpe Data. With nearly two decades of experience refining claims processes for carriers like Progressive and GEICO, Tom continues to make a profound impact on insurers’ claims fraud detection and management practices through his role guiding Carpe Data’s product development, strategy, and execution.
About Carpe Data – Using proprietary algorithms and proven AI, Carpe Data harnesses the power of emerging and alternative data for insurance carriers around the globe. Utilizing Carpe Data’s data, insurers gain deeper insight into risks and significantly enhance many aspects of the insurance life cycle, including underwriting, claims, and book assessment.
Let’s start…
FinTech Insights (FTI): Hi Tom, Welcome to the Fintech Top Voice Interview Series. Can you share a brief overview of your professional journey so far?
Tom: I have nearly two decades of experience in the insurance industry, focusing on refining claims processes for insurers like Progressive and Geico. I took this carrier-specific claims and fraud detection experience over to Carpe Data, where I now lead our claims product portfolio as the Vice President of Product – Claims. I’m excited to be on the technology side of insurance helping support the future of the industry in identifying fraud, streamlining claims management, and ultimately enhancing better, data-driven decision-making.
FTI: How can Carpe Data’s solutions help insurers balance innovation with ROI?
Tom: As slow-to-evolve industries like insurance fear being left behind in the current age of AI innovation, many companies are jumping onto the AI hype bandwagon too fast, sacrificing financial resources without any real strategy in place.
Although AI is a powerful tool for insurance advancement, surface-level adoption does not translate these investments into tangible gains. At Carpe Data, we ensure that our solutions are built with insurance use cases in mind and that new innovations deliver immediately measurable ROI — not focused on delivering more innovation but on delivering the right kind of innovation.
Innovation for its own sake is a mentality that has burned the insurance industry in years past and now it’s critical that investments in new tools and technology are conducted intentionally for concrete business problems.
Fintech Insights: Fintech Top Voice: Interview with Carolyn Duby, Field CTO at Cloudera
FTI: How do you see the role of alternative data evolving in the insurance industry over the next five years
Tom: Online data has grown increasingly relevant as much of our daily life is now lived digitally. It’s prudent for the insurance industry to not only tap into this growing source of alternative data but to also explore solutions that can effectively distill these oceans of unstructured information into actionable insights for better, faster, and more informed decisions.
Despite being one of the most data-hungry sectors in the world, many insurers are bogged down by outdated, limited data sets combined with time-intensive processes that often fail to gather the necessary insights to drive meaningful decisions. By scraping the internet for publicly available online content such as social media and Google reviews, insurers can derive unique insights that will allow them to streamline operations, combat the burgeoning problem of fraud, and get to the ground truth of what’s really happening in the claims process across their entire portfolio.
Risks are growing more complex, fraud is becoming an increasingly prevalent issue, and a diminishing talent pool is leaving claims adjusters exhausted and unmotivated. Intelligent insights fueled by alternative data provide a more comprehensive view of risk that empowers insurers to navigate a tumultuous industry landscape with veracity and confidence.
FTI: How do Carpe Data’s AI-driven tools transform traditional fraud detection methods, and what measurable impacts have you seen on claims processes?
Tom: Insurance fraud in America costs $308.6B annually, yet most insurers today only review 5% of their open claims for potential fraud, exclusively relying on an adjuster’s “intuition” to flag fraudulent activity once it’s already happened.
Proactive, automated analytical tools like Carpe Data’s Online Injury Alerts provide insurers with a complete picture of all the people and businesses they interact with every day, across their entire book of business. It’s critical to holistically understand risk and spot potential fraud before it occurs and we’ve found that 7% of carriers’ injury claims had a fraudulent or risky behavior they should be alerted to.
However, the solution is not to inundate overburdened and understaffed carriers with even more data and tedious research tasks. Carpe Data’s clients shared that – of the 7% of injury claims flagged – over 80% of the associated alerts they received were actionable. These evidence-based insights streamline and clarify decision-making to reduce fraudulent claim payouts, avoid costly litigation, and expedite relief to legitimate claims in less time.
In addition, nuclear verdicts, defined as a claims settlement of at least $10 million, reached an all-time high in 2023 with a median average value of $44 million, up from a low of $21 million in 2020. As attorney-driven inflation has skyrocketed the severity of fraudulent injury claims, evidentiary-based alerts can help insurers defend against exaggerated settlements and high litigation costs while expediting relief to legitimate policyholders.
FTI: How would you handle biases in AI models, particularly when applied to sensitive industries like insurance?
Tom: AI must be used as a way to augment rather than replace humans. By combining the best of both worlds — human decision-making, empathy, and expertise with automation and enhanced visibility into the overall nature of claims — sensitive industries like insurance can succeed in the modern age while still acting accountably and ethically.
At Carpe Data we believe AI models shouldn’t operate as “black boxes” where outputs lack explanation or context. For example, assigning a policyholder a number on a scale of 1-10 to quantify how likely they’re committing fraud is vague and still doesn’t inform a carrier on how they should proceed.
When AI models make decisions in isolation, it can lead to significant problems. We want to ensure that AI-powered claims solutions provide evidence-based alerts rather than abstract and nebulous assessments. Then, the human adjuster is equipped with the necessary information to make the final decision quickly and effectively based on transparent facts rather than mysterious algorithms.
FTI: What metrics would you recommend to assess the success of integrating AI into the insurance life cycle?
Tom: Operational efficiency gains
- Adjusters typically spend 15-30 minutes or more per claim investigating potential fraud online, depending on its complexity. A streamlined process can save adjusters hundreds of hours and a lot of headaches by freeing them from grueling, manual processes. These time savings directly translate into a happier, more efficient, and more productive workforce.
- The purpose of AI isn’t to blind companies with a blizzard of additional data, it’s about cutting through the noise to deliver valuable insights. False positives and superfluous information can detract rather than expedite the work, covering the path forward under mountains of snow rather than serving as a beacon of warmth.
Speed and ease of implementation
- AI tools can achieve many results, but if those outputs require an immense time investment to implement, a complete overhaul of existing systems, or ultimately add more work to employees due to insufficient training, then that is not a successful or strategic approach.
- Carpe Data’s Claims Solutions can be up and running in 2 weeks, accompanied by personalized training for end users, and start saving companies money within the first month.
Direct ROI and number of fraudulent claims detected
- Each fraudulent claim, on average, costs insurers $3,500. Tools that proactively detect fraud as early as First Notification of Loss (FNOL) can immediately translate into dollars and time saved by avoiding fraudulent payouts and lengthy litigation. Evaluating insurance reserve depletion year over year can illustrate the success of these AI tools and, if harnessed well, a carrier with 250,000 open injury claims could anticipate saving more than $60M annually.
Fintech Insights: Fintech Top Voice: Interview with Arcady Lapiro, CEO & Founder at Agora Financial Technologies
FTI: What are the key challenges in integrating emerging data sources into underwriting and claims processes?
Two of the primary challenges with integrating emerging data sources are sourcing high-quality, robust data, and ensuring it can be properly implemented into workflows. Many of the existing data sets insurers use are both small and outdated.
Organizations that rely on data pulled from millions of online sources and updated every month benefit from a much more comprehensive view of a policyholder and the nature of their claim. This fuller picture of loss facts and evidence equips insurers with greater confidence and accuracy for real-time decisions resulting in swifter claim resolutions and happier customers.
This alternative data also needs to be easily integrable. Having all of that data is useless if it needs to be pored over, analyzed, and formatted before it can be distilled into actionable insights. Having this data structured and easily accessible through existing workflows is the final cincher for meaningfully harnessing it.
FTI: In your experience, what’s the most critical step in building a seamless pipeline for alternative data integration?
Ensuring data quality and relevance. Make sure you’re pulling from the right sources and ensuring that you’re delivering the right insights, not just more information.
Insurers aren’t seeking more data to complicate decision-making, they are looking for insights elicited by AI-powered analytics that inform action with greater efficiency and confidence. Overloading insurers with noise and extraneous alerts detracts from the solution rather than realizing it.
FTI: What are the biggest challenges insurers face with decision-making and how can high-quality, curated data address these issues?
Adjusters are overwhelmed by massive amounts of data- a great deal of it being irrelevant, outdated, and unreliable — making it difficult to act confidently. Because 30% of people under 45 don’t view insurance fraud as a crime, this issue will only become a larger weight for insurers to lift. Scarce adjuster teams relying on time-intensive methods to parse valuable insights from oceans of low-quality, unstructured data results in greater decision fatigue, more frequent and severe fraudulent payouts, and slower claims resolution.
It’s important to note that better data does not mean more data. High-quality, curated data that is easily translated into actionable, targeted insights is key for insurers to step away from traditional, reactive approaches and move forward with confidence to process claims faster, identify patterns of fraudulent activity, and ultimately support policyholders with lower premiums and better customer service.
FTI: A fresh tip on FinTech that you would like to give our readers.
Don’t just focus on exciting innovation, focus on strategic and practical innovation.
Cutting-edge technology can often be rooted in lofty promises with unproven results. The investments your company makes must be rooted in ease of implementation, operational efficiency, and direct ROI. That way, you will meaningfully support your workforce, drive results for your customers, and address the most pressing business problems now and in the future.
FTI: What are your predictions for 2025 in the FinTech domain?
AI: Offerings will focus less on breaking the mold and more on delivering real-world value.
Fraud: Will become a central issue for insurers as the staggering cost of fraud continues to outpace even natural catastrophe losses by more than 200% in America alone. Also, the shifting consumer views on insurance fraud as a crime will only exacerbate people’s willingness to commit it.
Carrier collaboration: As fraudulent activity grows, insurers will begin to rely more on “the network effect” where carriers can work together to analyze fraud and risk that was able to breach their individual defenses. By sharing intelligence, pattern recognition, and data, insurers will be able to more effectively identify and disrupt organized fraudulent activity. This rising tide that lifts all ships will result in a sweeping reduction of fraud that trickles down into fairer assessments and reduced premiums.
Fintech Insights: Fintech Top Voice: Interview with Tendü Yogurtçu, CTO at Precisely
Thank you Tom for sharing your Fintech comments. We look forward to having Carpe Data again at the Fintech top voice series.
