Watch Indico Data CEO Tom Wilde step in as co-host alongside Michelle Gouveia, VP at Sandbox Insurtech Ventures, in season 2 episode 14 of Unstructured Unlocked with Louis DiModugno, Global Chief Data Officer at Verisk.
Michelle Gouveia: Hey everybody. Welcome to another episode of Unstructured Unlocked. I’m Michelle Gouveia.
Tom Wilde: And I’m co-host Tom Wilde.
Michelle Gouveia: We are very excited to welcome today’s guest, Louis DiModugno. Some of you may remember him from a previous episode, and weâre really happy to have him back.
Louis is now the Global Chief Data Officer at Verisk. Welcome back, Louis! Thanks for joining us again.
Louis DiModugno: Thanks so much for having me. I’m excited to be here.
Michelle Gouveia: Really excited to have you back! Last time, we spent a lot of time talking about data governance and data within the four walls of an insurance carrier. Today, Iâd love to hear how your perspective has evolved now that youâre working on the other sideâas a partner and vendor to carriers.
Before we dive into all of that, can you share a little bit about your role today and what youâve been up to since the last time you joined us?
Louis DiModugno: Sure. As you mentioned, I recently joined Verisk as Global Chief Data Officer. Whatâs interesting about this role is that itâs aligned with the IT side of the house, which makes sense given our focus.
When I first joined, I saw many challengesâbut Iâve since started seeing them as opportunities. My role focuses on helping carriers aggregate, normalize, and integrate their data, performing entity resolution and master data management.
Verisk provides data solutions that span the entire insurance value chain, from pricing and underwriting to claims and extreme event analysis. What makes us unique is that we donât just work at a single point in the processâwe provide insights at every step of the way.
Tom Wilde: The role of a Chief Data Officer varies significantly from company to company. How does your mandate at Verisk differ from your experience inside an insurance carrier?
Louis DiModugno: A key difference is that at Verisk, weâre working further downstream from the initial data capture.
In the insurance industry, data capture primarily happens at the agency and broker level. That data may be entered directly into a system, or in some cases, itâs still written down with a stubby pencil. One of the biggest ongoing challenges in this space is ensuring data is accurately captured from the start.
Once that data reaches us at Verisk, our job is to normalize it, resolve discrepancies, and track how policyholders move through the insurance ecosystem over time. We analyze their exposure to products, claims history, and evolving needsâwhether personal or commercial.
By looking at this data over time, we can identify trends, understand customer movements, and provide carriers with deeper insights into industry shifts.
Michelle Gouveia: I think we actually talked about this last time and compared it to a game of telephone. The more entities that touch the data, the more risk there is of changes, misinterpretations, or even unintentional manipulation.
Verisk gets data in two main waysâdirect contributions from carriers and data aggregated from various sources. How do you think about governance and compliance when managing data from these different sources? And what should carriers keep in mind when sharing data with Verisk?
Louis DiModugno: Our priority is always ensuring the highest quality data.
When we receive data from carriers, we compare it to previous submissions to track changes and maintain consistency. We also perform entity resolution to understand how policyholders move through the insurance ecosystem.
One key point is that as we move further into AI and generative AI, data quality becomes even more critical. The accuracy of these models is only as good as the quality of the input data.
For carriers, the focus should be on capturing accurate data from the beginning. Many carriers have their own initiatives for creating “gold standard” data within their policy admin systems. We hope that by the time data reaches us, it has already undergone internal quality checksâallowing us to focus on normalization and integration rather than fixing inaccuracies.
Tom Wilde: The insurance industry has long struggled with standardization. In commercial and specialty insurance, especially, it feels like a Tower of Babel situationâso many different data formats, sources, and variations.
Do you see a future where this improves, or is the reality that this complexity will persist? Does that mean the role of companies like Veriskâhelping make sense of all this dataâremains critical?
Louis DiModugno: One of the advantages we have at Verisk is that, as a data aggregator, weâve been able to set standards.
We dictate to the industry the format, structure, and frequency with which we receive data from carriers. This level of standardization has been key to ensuring consistency.
Thatâs not always the case in every part of the industry. For example, when I worked in reinsurance, the data we received varied significantly because each carrier sent it in a different format. Sometimes, it was even handwritten on paper!
What helps maintain consistency is our partnerships with policy admin system providersâcompanies like Guidewire, Duck Creek, and EIS. These platforms are where much of the industryâs data originates, so by working closely with them, we can ensure a higher degree of standardization.
Michelle Gouveia: Within Verisk, how do you approach data governance and compliance given the vast amount of data coming in from multiple sources? And how do you position the strength of your governance as part of the value proposition for Veriskâs solutions?
Louis DiModugno: Data governance is a huge focus for us. Some companies do the bare minimum to comply with regulations, but at Verisk, weâve taken a much more proactive approach.
Weâve gone beyond traditional governance frameworks and implemented a data observability approach. This includes monitoring:
Because we have contractual obligations with carriers, we have strict controls on data usage. Every time a data scientist or underwriter wants to use specific data, they must request access through a governance board.
Weâve built a controlled environment where we can track exactly who accesses what data, how often, and for what purpose. This ensures compliance with both regulations and carrier agreements.
Additionally, we look at the cost of data managementâunderstanding the return on investment in improving data quality and ensuring that our data efforts drive tangible business value.
Tom Wilde: Has generative AI changed your mission, or is it more of an evolution of traditional AI challenges? AI has always been about garbage in, garbage out, and Gen AI seems to amplify thatâmeaning mistakes can now happen at scale and with more speed.
Louis DiModugno: Generative AI has created significant efficiency gains for us, but weâve implemented safeguards to ensure accuracy.
For example, we use a retrieval-augmented generation (RAG) architecture in our AI models. This means that whenever the AI provides an answer, it also includes references showing where the information came from.
A practical use case for us is reviewing legal contracts. Instead of having a legal professional spend hours manually combing through documents, we can use AI to quickly find relevant clauses while providing citations.
The key is ensuring that AI is used responsibly, with human oversight to verify outputs.
Michelle Gouveia: What are your thoughts on how AI will change data regulations? Weâve seen new guidance from state insurance departments on the use of AIâhow do you see this evolving?
Louis DiModugno: Regulation is evolving rapidly, especially in insurance.
With 50+ regulatory environments in the U.S. alone, keeping up with compliance changes is a major challenge. AI can help by tracking and summarizing these changes, allowing companies to adapt more efficiently.
Another key area is fraud detectionâAI is already being used to analyze claims data, images, and adjuster notes to detect patterns of fraudulent behavior. The ability to process and interpret unstructured data at scale is creating new opportunities for risk management.
Tom Wilde: Looking ahead five years, whatâs your vision for how data and AI will transform the insurance industry?
Louis DiModugno: My biggest hope is that insurance becomes far easier for consumers to understand.
Right now, insurance products are complex and confusing. Many people donât fully understand their coverage, exclusions, or how their claims impact premiums.
If we can use AI and data-driven insights to democratize insurance knowledge, it will be a huge win. Consumers will make better decisions, access the right products, and ultimately see more value from their policies.
Tom Wilde: Thatâs a great note to end on. Louis, thanks so much for joining us again!
Michelle Gouveia: We really appreciate it, Louis.
Louis DiModugno: Always a pleasure. Thanks for having me!
Check out the full Unstructured Unlocked podcast on your favorite platform, including: