1.888.958.5846
info@gradientai.com
From the Floor at NAPEO’s Risk Management Workshop: Industry Panelists Share Insights on AI in Workers’ Comp
NAPEO's Risk Management Workshop has become one of the largest gatherings of PEO risk management professionals, carriers, brokers, and agents. This year’s programming centered around the ever-changing topic of artificial intelligence and how it impacts the world of PEO risk management.
Sessions included a workshop on the role of AI in insurance underwriting, which was moderated by Gradient AI’s Chase Pettus, Vice President of Sales. Pettus was joined by panelists, Frank Huang, Managing Director and Consulting Actuary,
Davies Group North America; Julie Cirillo, Chief Risk Officer at
Engage PEO, and Justin Rowley, Vice President Risk Management,
Helpside.
Insights from the Panel: “AI & Risk - The Role of AI in Insurance Underwriting”
The panelists provided a unique view of AI in the Workers’ Comp industry, as well as shared their first-hand experience with the use of AI in their own companies.
The engaging session focused on key discussion points, including:
1. What is driving the use of AI for PEOs and in the Workers’ Comp Market?
2. What challenges are companies trying to solve for with AI?
3. What is the Impact of AI in Workers’ Compensation?
4. What results have you seen since incorporating AI?
5. Why are some companies slow to implement AI as part of their operations?
6. What advice do the panelists have for those considering AI?
Below, we delve deeper into these discussion points.
1. What is driving AI for PEOs and in the Workers’ Comp Market?
There are many reasons for using AI in insurance, including keeping ahead of the competition, improved efficiency, and mitigating a lack of internal data. The panelists agreed that the motivation to consider using AI in underwriting varies from overcoming staff turnover in underwriting and claims, bolstering departments short on staff and experience, to the desire to expand into other territories and markets.
However, while the reasons for using AI may differ, Frank Huang of Davies Group North America, whose experience ranges across P&C, with a particular focus in workers’ compensation, general liability, and employment practices liability, identified one key underlying factor that is common to all: “The overarching theme is profitability. Not only are companies always looking to improve the bottom line, but they want to make sure they are not being adversely selected against. Essentially, incorporating AI helps them avoid being left behind.”
2. What challenges are companies trying to solve with AI?
While Workers’ Comp is still in a soft market, it can be a challenging market for some insurers and risk takers. Decreasing rates and ever-changing claim costs can lead to severe pricing pressure. In addition, it can be difficult to get the clearest picture of risk.
Helpside, a Utah-based PEO providing solutions for payroll, employee benefits, human resources, and work comp solutions, was an early adopter of AI in workers’ comp - incorporating AI about seven years ago. “Initially, we were looking to add more certainty to our underwriting process,” said Helpside’s Justin Rowley during the panel discussion. “We wanted to add another piece of data to the model we already used, especially in states we weren’t familiar with and for companies that didn’t have loss information, such as start-ups. AI provided that additional data.”
Engage PEO is leveraging AI in its underwriting solution to enhance its workers’ comp underwriting process and fuel its business expansion as part of Engage PEO’s broader AI strategy to facilitate operational efficiency and improve results. Julie Cirillo echoed Frank Huang’s sentiments as to why Engage PEO incorporated AI and what challenges they were looking to solve.
“We really wanted to be profitable,” said Cirillo. “And I wanted to find a way, as Engage PEO grew across the country and moved into new territories or industries, to accurately predict the loss for each prospect so we can price accordingly. For example, California can be a challenging state for workers’ compensation, and you had better have a tool that helps you understand the difference between underwriting a prospect in Los Angeles versus one in Sacramento.”
3. What impact has AI had on Workers’ Comp?
From a broad market view, Frank Huang shared that claims operations saw an early impact when simpler claims models came out about twenty years ago, that allowed better claims triage.
“With AI, you could have 100 cases come in and you will be able to identify and triage the cases that are most likely to incur higher costs, result in potential litigation, or involve greater claims management,” said Huang. “AI models also impact claims because of the year-round and heavily manual processes that have historically been used to manage claims, aggravated by the turnover in experienced complex claims adjusters.”
As AI Underwriting models evolved, Huang added the important impact on proactive risk management. “As I noted earlier, AI models can help you identify the potential for larger, more lengthy claims and support better claims management, but with the underwriting models available, you are now asking, ‘do we want to allow this business in the door and are we sure have the right price points?”
4. What results have you seen since incorporating AI?
Providing the perspective of a PEO’s experience first-hand, Justin Rowley shared that AI has not only been helpful to Helpside in pricing new clients, but also in their renewal business. “The biggest impact AI has had on us, has been increased certainty in the underwriting process,” said Rowley. “We still use loss runs and an actuarial model when possible. I have built a dashboard that allows me to see the actuarial predicted losses as well as the AI predicted losses and we make decisions from there.”
Rowley also added three key additional benefits of utilizing AI in workers’ comp for Helpside, including:
- Greater insights for prospects with gaps in data
- Quicker underwriting turnaround times
- Increased underwriting efficiency
Julie Cirillo said that Engage needed to create a sustainable underwriting model to support that growth she spoke about earlier in the program. AI has enabled Cirillo to build an underwriting team that can:
- Accurately evaluate risk
- Onboard and empower less experienced underwriters
- Support growth while improving efficiency
“I needed to know that regardless of which underwriter looked at the potential risk of a prospect, we would get consistent results,” said Cirillo.
Not only is AI significantly enhancing Engage PEO’s underwriting process, but it is also being used to support the company’s M&A activities by evaluating segments of risk within a whole book of business. Cirillo added, “This solution has assisted in integrating our acquired companies, allowing for fast and accurate review of both the portfolio and individual clients to ensure profitability is maintained with the transition to our risk-bearing program.”
5. Why are some companies slow to implement AI as part of their operations?
“I think possibly fear or lack of knowledge,” said Frank Huang, who observed the slow adoption practices are not limited to the PEO space - but to the broader insurance and carrier industries too.
Huang did note that there are plenty of smaller firms not using AI that are doing fine. “However, Huang added,” as these smaller firms seek to grow and gain greater efficiency – coupled with the increase in AI usage in workers’ compensation, the potential consequence of not adopting AI does theoretically grow.”
The audience provided feedback on potential barriers as well, expressing concerns around cost and change to underwriting teams experience and habits.
“An AI model may not seem cheap,” said Julie Cirillo. “But you have to trust in your investment. If I could show you the profitability for Engage PEO on the Workers’ Comp line, it is significant. And it is significantly improved year over year. I credit that to our AI model that we are selecting the right risk and pricing it correctly. The investment has paid for itself.”
6. What advice do the panelists have for those considering AI?
“It is helpful to know what you are you trying to solve,” said Frank Huang. “For instance, if it is a focus on profitability or overcoming lack of experience than develop a model to support this. If you are looking to save five points of return to justify the investment, map out how you get there and how the tool supports that.”
“We recently had discussions with potential investors and partners and AI was a big part of those conversations,” said Julie Cirillo. “What is the future for your PEO? If you are looking to grow and sell or recapitalize, I think you will find AI is going to be a critical part of that growth.”
Cirillo added that the human element does not go away with AI. “I would never let a model just kick a quote to sales. Our underwriting team is always going to review the price range that comes out of the model. I would say that 90% of the time we price within the range but another 10% of the time we use underwriter discretion and manage approvals accordingly.”
“I think we all want to work towards better profitability,” said Justin Rowley. “But keep in mind, AI does not replace solid underwriting practices or business practices. Utilize AI to enhance these practices.”
Learn More
Read the press release and discover how Engage PEO utilized Gradient AI’s solution to increase risk appetite, drive profitability, and spur rapid expansion. Read more about Gradient AI’s Workers’ Comp solutions, and solutions for PEOs in Group Health.
About NAPEO
The National Association of Professional Employer Organizations is The Voice of the PEO IndustryTM. PEOs provide payroll, benefits, regulatory compliance assistance, and other HR services to small and mid-sized companies. NAPEO's mission is to grow, promote, and support the PEO industry. NAPEO advocates for the interests of PEOs at all levels of government, and provides robust member resources, vital networking and referrals, relevant and timely education, and effective public relations and marketing support.