Artificial intelligence has become a strategic priority across the insurance sector. At Skyward Specialty Insurance, the emphasis is clear: strengthen underwriting at every point in the cycle.
For Dan Bodnar (pictured), CIO of the Houston-based specialty carrier, the underwriting function sits at the center of long-term value creation. Claims tools and back-office efficiency remain important, but the underwriting cycle commands the greatest attention.
“For us, the underwriter is the center of everything, and the underwriting cycle is our primary focus,” Bodnar said.
That focus translates into a multi-year transformation effort spanning submission ingestion, data augmentation, AI-assisted underwriting and enterprise analytics. The strategy combines internal builds with selective external partnerships, underpinned by a deliberate approach to change management.
Skyward has targeted the earliest stage of the underwriting workflow – submission intake – as a priority.
The insurer has rolled out a company-wide ingestion tool built internally. The platform captures inbound emails, formats submissions and preloads structured data into documentation, policy administration and analytics systems.
“It takes the human out of the heavy lift of ingesting and structuring emails with a bunch of attachments,” Bodnar said.
Ingestion, he said, represents a significant investment area, particularly as submission volumes and data complexity increase across specialty lines.
Alongside ingestion, the company has deepened its data augmentation capabilities. External data sources are layered with internal insights to support underwriting decisions and portfolio management.
Behind the scenes, analytics, business intelligence, modeling teams and actuaries are connected through a shared enterprise data architecture. Skyward’s BI platform – branded internally as SkyBI and built on Microsoft Power BI – has been in place for roughly three years, with ongoing enhancements as new data sets are integrated.
More recently, the insurer has incorporated agentic AI over its enterprise data warehouse. The goal is to accelerate access and insight for business leaders.
“You can now essentially speak reports into existence and run ad hoc deep dives before you ever need to call someone for help writing a report - something that used to take weeks,” Bodnar said. The capability, he said, is designed to reduce friction and shorten the time between question and answer.
Externally, one of Skyward’s more visible partnerships is with Sixfold, supporting what the company describes internally as “AI-assisted underwriting.”
The internal tagline – “bionic underwriting” – reflects a deliberate positioning.
“It’s important for people to understand – especially where we sit with complex risks – that AI is only going to assist underwriters, not replace them,” Bodnar said.
The Sixfold partnership is structured to keep the human in the loop while improving speed and depth of analysis. In complex specialty segments, Bodnar argues that judgment remains central, even as AI tools enhance productivity and pattern recognition.
Skyward’s broader technology strategy blends build and buy. Bodnar describes himself as “a ‘buy’ person,” preferring to partner externally unless a capability represents a durable differentiator.
Yet, he acknowledges that vendor selection has become more difficult.
“I don’t think it has gotten easier to choose external partners; I think it has gotten harder,” he said, particularly given the relative immaturity of some AI offerings and uncertainty around outcomes.
Ingestion illustrates the shifting calculus. What once appeared to be a strong differentiator is moving toward commoditization.
“It’s gotten so much better over just the last 18 months,” Bodnar said, noting that even Excel spreadsheets could now be reliably ingested if they were not overly complex.
With use cases multiplying, prioritization has become a discipline.
Skyward evaluates initiatives against three value-creation metrics: revenue impact, risk management improvement and efficiency gains.
“We have a lot of ideas, so we ask: does it drive revenue? Does it help us manage risk better? Or is it an efficiency play?” Bodnar said.
Speed also plays a role. Demonstrating working prototypes within weeks, rather than debating concepts for months, helps secure internal buy-in.
“You can talk about something for three months, but if you can show people something in two weeks – and they can see where they could be in three months – you’ve got an engaged partner at that point,” he said.
Change management, however, remains critical. Bodnar acknowledges that some employees felt uncertain or threatened by AI. Early initiatives are therefore focused on automating routine tasks rather than encroaching on judgement-based underwriting decisions.
“Everyone wants to get rid of the grunt work they do every day and wish they could get off their desk,” he said. “What you choose to do first is important. It should be rooted in ROI.”
Executive sponsorship underpins the programme. Bodnar said innovation would stall without visible and consistent leadership support.
“You’re not going to get far without executive and senior leadership support,” he said, describing experimentation and failing fast as embedded in the company’s culture.
He views AI as a general-purpose technology – comparable to electricity or the steam engine – whose full implications were still unfolding. As a result, innovation flows both top-down and bottom-up.
Strategic initiatives such as ingestion and analytics are enterprise-led. At the same time, the company captures “desktop-up” ideas from employees experimenting with emerging tools.
“Some of the most leverageable ideas come from people doing their jobs and playing with tools like ChatGPT and realizing they can do things a lot better,” Bodnar said.
At a recent senior leadership meeting in Orlando, 100 leaders generated 84 AI use cases sourced from across the organization. Those ideas are entering a formal vetting and prioritization process.
For Bodnar, the differentiator in an AI-driven market is not simply technical capability, but adaptability.
“Chaos is the new normal,” he said. Companies that build systematic processes to capture, vet and execute innovation will be better positioned to sustain underwriting performance and portfolio resilience over time.
