How Bill Devine, Former VP at Travelers, Sees AI Revolutionizing Insurance Underwriting

Things are changing in the insurance industry—AI and automated decisioning are reshaping workflows, elevating customer experiences, and streamlining operations. These advancements were at the center of a recent episode of Unstructured Unlocked, featuring Former VP of Travelers, Bill Devine. The conversation highlighted the challenges facing traditional underwriting processes and the immense potential AI holds for carriers navigating this complex landscape.

Listen to the full podcast here: Unstructured Unlocked season 2, episode 16 with Bill Devine, Former Travelers Insurance VP

The hurdles of traditional underwriting

Underwriting, the backbone of the insurance industry, has historically relied on manual workflows. Devine described the importance of each underwriter’s expertise: “Every underwriter’s desk was like a cottage industry.” He emphasized how underwriting has always relied heavily on individual experience and information. This artisan-like approach is incredibly important and must be respected and upheld; however, this method struggles to scale by itself in today’s data-driven environment.

Devine also highlighted the legacy challenges insurers face: “Part of our legacy culture as an industry that we’ve developed is a predilection for building things ourselves.” While insurers have been at the forefront of adopting technologies like SaaS and cloud platforms in the past, this self-reliant culture has left many companies with outdated legacy systems ill-equipped to handle modern demands or adopt technology like AI.

Another issue: the fragmented nature of underwriting workflows often leads to delays, inaccuracies, and inefficiencies. Insurers must sift through a “river of submissions,” prioritizing high-value risks while processing huge volumes of data, much of it unstructured. In this environment, traditional methods of decision-making fall short, especially as customer expectations rise and regulatory pressures mount.

AI and the decision era: A new paradigm

The insurance industry is entering the “Decision Era,” where AI transforms data into actionable insights, enabling faster, more accurate decisioning across the policy lifecycle. Tom Wilde, CEO of Indico, underscored this shift: “We’re at a moment in time where compute, data, [and] AI allow us suddenly to unlock and capture the ROI.”

Wilde explained that AI allows insurers to take the millions of documents sitting in file stores—traditionally dark data—and turn them into decision-ready assets. This capability fundamentally changes how insurers approach underwriting, shifting from reactive processes to proactive decisioning that is faster, more precise, and scalable.

Related content: How underwriting triage powered by AI improves risk management

Improving decisioning through collaboration

While AI automates many aspects of underwriting, human expertise remains indispensable. Devine stressed the importance of collaboration: “At the end of the day, insurance is a transaction handled between a broker and an underwriter… So ‘human in the loop’ is an incredibly important part of, I think, the entirety of this [process].”

This human-AI partnership ensures that automated tools augment rather than replace underwriters. AI systems can handle data aggregation, risk scoring, and triage, but complex, heterogeneous risks often require human judgment and contextual understanding. As Wilde noted, the future of AI in insurance lies in “lowercase ‘a’ automation” (as opposed to “uppercase ‘A’ Automation”), where technology accelerates processes without compromising accuracy or explainability.

Explainability has become a critical component of AI deployment in insurance. Regulators increasingly demand transparency in AI-driven decisioning, requiring insurers to provide clear justifications for their actions. Wilde emphasized Indico’s approach to this challenge, highlighting how they built explainability into the AI development process early to ensure that every step in the decision supply chain is traceable and auditable.

The decision supply chain: A roadmap for AI-driven workflows

The concept of the decision supply chain provides a powerful framework for understanding how AI can transform underwriting and other insurance processes. Similar to a manufacturing supply chain, the decision supply chain involves sequential steps that convert raw materials—in this case, data—into actionable insights. These steps take raw data through six steps leading to an ultimate decision: source, ingestion, extraction, validation, application, and context. By properly executing each phase, insurers can efficiently verify their data and make it decision-ready, turning it into valuable inputs for underwriting and risk assessment.

Wilde emphasized the importance of this approach, stating that the decision supply chain ensures explainability and transparency at every stage. For example, AI systems can extract insights from unstructured documents, validate them against underwriting guidelines, and prioritize submissions based on risk appetite. This streamlined process not only accelerates workflows but also enhances the accuracy and consistency of decisions, positioning insurers to thrive in a competitive and data-driven marketplace. By adopting the decision supply chain metaphor, insurers can reimagine their operations as cohesive, AI-powered ecosystems.

Transforming the underwriting process with intelligent solutions

Indico’s intelligent solutions, including its new clearance and triage platform, exemplify how AI can streamline underwriting workflows. The platform automates key steps, from submission intake to document classification and risk scoring, reducing processing times from days to hours. By integrating AI tools like generative AI for summarization and Agentic AI for decision-making, the platform empowers insurers to quote more efficiently and focus resources where they matter most.

For example, Wilde shared how Indico’s new underwriting clearance and triage platform enables insurers to adapt to changing market conditions: “…In a hard market, it’s just not as critical to be able to efficiently understand the river of submissions you’re getting… In a soft market, that changes dramatically. You need to be able to access that full river of submissions much more efficiently because you’re going to have to assess and quote a much larger percentage of that submission river to hit the same goals that you had.” He continued, saying that insurers have “got to be able to leverage [their] underwriting resources more effectively, but you can’t suddenly double the number of underwriters you have to get at the fact that you’re going to have to triage and quote more of the submissions. You need to give them tools, co-pilots, bionic arms to take advantage of that and leverage the existing team you have.” Thus, AI-based underwriting tools.

Data as the foundation for AI success

The discussion returned multiple times to the fact that data is what drives AI’s effectiveness. Devine observed that while the insurance industry has amassed vast amounts of data over decades, much of it remains underutilized. “We have been in the data and the information game for so long, we as an industry have potentially massive accumulations of data at this point. I think the challenge is how much of that data is actionable, how much of that data is usable? …We have unbelievable amounts of data because making data-driven decisions has been in the DNA of the industry. But I think the flip side… is that just means we have massive amounts of data, and to be able to get through it all intelligently and to use it is something that’s a challenge now.”

To address this, insurers must focus on data quality and accessibility. Wilde emphasized the importance of “good ground truth” for AI systems, noting that “AI is still very much a ‘garbage-in, garbage-out’ equation.” Insurers that invest in data cleansing and integration are better positioned to unlock the full potential of AI-enhanced decisioning.

Moreover, unstructured data—long considered an untapped resource—is now a critical asset. AI technologies like generative AI can extract insights from these data sources, enabling longitudinal analysis across portfolios rather than isolated risk assessments. This capability allows insurers to identify patterns, optimize pricing strategies, and improve portfolio management.

Related content: AI-enhanced decisioning: Transforming the insurance submission clearance process

The road ahead for AI in insurance

As the insurance industry continues to embrace AI, the potential applications extend far beyond underwriting. From fraud detection to claims processing, AI-driven solutions promise to revolutionize every stage of the policy lifecycle. However, this transformation requires careful planning, robust governance, and a commitment to ethical decisioning.

Devine offered a perspective on the industry’s next steps: “Carrier companies should think more in the terms of consumption and less in terms of development.” The focus for carriers should be on leveraging AI tools to advance business goals rather than trying to become tech companies themselves.

Wilde echoed this sentiment, highlighting the shift toward “service as a software” models where insurers can license specialized solutions tailored to their needs. By partnering with technology providers like Indico, carriers can accelerate their digital transformation while maintaining focus on their core competencies.

Unlocking the future of insurance with Indico

This episode of the Unstructured Unlocked podcast underscored the critical role AI plays in shaping the future of insurance. By addressing the challenges of traditional underwriting, enhancing collaboration between humans and machines, and leveraging data as a strategic asset, AI-powered decisioning sets the stage for a new era of efficiency and innovation.

Indico’s intelligent solutions are at the forefront of this transformation, enabling insurers to navigate complexity, improve accuracy, and deliver better outcomes for policyholders. As Wilde aptly summarized, “The rise of AI has shown a really bright light on the need to have excellent data to be able to take advantage of AI and ultimately lead to what we think is the punchline, which is better decisions.”

To keep up with the latest in AI and insurance, you can subscribe to Unstructured Unlocked on your favorite platform:

• Apple Podcasts
• Spotify
• Amazon Music

Subscribe to our LinkedIn newsletter.

Frequently asked questions

• How are insurers managing the potential job displacement that AI could bring to underwriting roles, and what steps are being taken to retrain or upskill underwriters for this new era?
  The blog mentions the importance of human expertise in underwriting and the human-AI partnership but doesn’t address the implications for underwriters whose roles may change due to automation. Insurers are mitigating job displacement by focusing on reskilling initiatives, enabling underwriters to work alongside AI tools as strategic analysts rather than data processors. Training programs are being developed to enhance underwriters’ ability to interpret AI-generated insights and apply them to complex risk assessments. By redefining roles to emphasize decision-making and customer engagement, insurers ensure that their workforce remains relevant and empowered in the AI-driven landscape.
• What regulatory hurdles or challenges are insurers facing when implementing AI-driven underwriting tools, and how are they addressing these obstacles?
  While the blog touches on the importance of explainability for regulatory compliance, it doesn’t explore the specifics of navigating complex regulations in different jurisdictions. Insurers face challenges such as ensuring their AI systems align with privacy laws, anti-discrimination policies, and reporting requirements. To address these hurdles, companies are working closely with regulators and legal experts to design AI systems that meet compliance standards from the outset. This includes implementing audit trails, conducting regular bias testing, and maintaining transparency in AI decision-making processes. Collaborations between industry players and regulatory bodies are also paving the way for clearer guidelines on AI deployment in insurance.
• How does the shift to AI-powered underwriting impact smaller or regional insurers that may lack the resources to invest in advanced technologies?
  The article discusses the benefits of AI but doesn’t explore how smaller insurers might be affected by these industry-wide changes. Smaller insurers often face challenges in adopting AI due to limited budgets and technical expertise. To overcome these barriers, they are increasingly relying on partnerships with technology providers offering scalable, subscription-based AI solutions. By adopting service-as-a-software models, smaller insurers can integrate AI capabilities without the need for significant upfront investments. These partnerships also provide access to pre-trained AI models and support services, enabling smaller players to compete with larger carriers in terms of efficiency and accuracy.