Employers have experienced mixed results in recent years when using new technologies to improve the benefits open enrollment process. But the arrival of generative artificial intelligence and improved analytics software means 2023 may be the year technology provides a bigger payoff in helping employees make choices about benefits options, lightens HR professional's support burden and helps analyze enrollment data to improve future planning.
Benefits of Generative AI, Analytics Tools
Benefits open enrollment is an increasingly high-stakes process for organizations. In a 2022-2023 study by technology vendor isolved, 64 percent of respondents said a poor enrollment experience would cause them to look for new employment. That study also found more employees have come to expect technology will play an instrumental role in helping them make better benefits elections. For example, a significant number of respondents in the isolved study said the availability of a virtual assistant to guide them to the best benefits choices was an important option for them.
Technology analysts say one of the biggest payoffs employees will see from the use of generative AI during 2023 open enrollment is the improved functionality of chatbots they rely on to make benefits elections.
"Most of the chatbots before the arrival of GenAI were one-dimensional," said Rhonda Marcucci, vice president of innovation for Gallagher, an HR and benefits technology advisory firm. "But ChatGPT and other tools are making them smarter and expanding their functionality. Now a chatbot can go look at four benefits documents and summarize their key points for a user in a few sentences. That can be a big help in benefits decision support during open enrollment."
Experts say employees should see improved decision support as more benefits technology platforms incorporate GenAI into their chatbots. "One of the key innovations generative AI is facilitating is in natural language understanding technology," said Ron Hanscome, a research vice president specializing in HR technologies for Gartner. "It can generate much more human-like text and enables chatbots to process unstructured data more efficiently."
Businesssolver, a benefits technology provider in West Des Moines, Iowa, built ChatGPT into its virtual assistant Sofia.
ChatGPT enhances the natural language understanding (NLU) capabilities of the chatbot by helping it better decipher requests from employees and enables it to handle multiple questions in the same service interaction.
"As one example, an employee might ask the chatbot what their HSA balance is, followed by saying they need to add a dependent to their coverage," said Sony Sung-Chu, head of science and innovation at Businesssolver. "Before adding GenAI, our technology had a hard time discerning between the two inquiries made in one interaction. But now the chatbot can quickly determine there are two actions an employee wants to take, ask them which they want to handle first and then address both requests."
Jeff Robbins, managing director of outsourcing shared services for advisory firm WTW, said Microsoft's recent introduction of Azure Cognitive Service for Language allows chatbots built on the Microsoft Azure framework to better utilize large language model (LLM) capabilities.
"That enhances the bots' understanding of user inquiries and improves their ability to address user needs effectively," Robbins said.
He said robust decision support tools also can leverage employee data, claims history and preference profiles to suggest coverage options. For example, WTW has a "selection analysis" tool that can assess employee elections and provide personalized suggestions for workers to consider before they finalize their elections, Robbins said.
Sung-Chu said Businesssolver also is continually working to give its chatbot more human-like qualities, such as making it more empathetic and responsive to users.
"ChatGPT allows you to be more accurate in assessing the tone of an interaction," he said. "That might take the form of the chatbot offering a congratulations because an employee just graduated or showing empathy because someone just lost a family member."
Summarizing Benefits Documents and Content Generation
Generative AI also promises to aid the open enrollment process by making it easier to extract key points from benefits information, summarize customer call data and create important benefits content, experts say.
Businesssolver, for example, is using ChatGPT in its customer contact center to summarize incoming calls after they are automatically transcribed. That allows call center representatives to focus more on providing quality service during calls while ChatGPT summarizes and conducts sentiment analysis.
At WTW, customer call centers also are using generative AI to draft call summaries, Robbins said.
"The emergence of generative AI technology is revolutionizing the enrollment experience for clients, employees, consultants and administrators," Robbins said. "Another one of its primary applications is in content creation, significantly reducing the time and effort required to develop high-quality resources for employees throughout the enrollment process."
How AI Can Help Before, After Open Enrollment
New technologies also can help HR and benefits leaders make better decisions both before and after open enrollment. Prior to the start of enrollment, they can use AI to help make decisions on coverage design, costs and employee education campaigns for the upcoming year, experts said.
"AI-enabled technology, such as enhanced analytics and predictive modeling tools, assists in making informed decisions that align benefit designs with employee needs and company objectives," Robbins said.
After enrollment closes, organizations also can use AI to generate valuable data on employee behavior and elections, Robbins said, including using sentiment analysis tools to evaluate feedback that can improve the process for the next year.
"Data scientists can leverage AI and sophisticated analytics technology to extract powerful insights from this data," Robbins said. "By using population segmentation and sentiment analysis, employers can identify opportunities within their diverse employee population. AI and automation have significantly reduced the time and effort required to deliver this information, with real-time availability in many cases, compared to the months it used to take."
Dave Zielinski is principal of Skiwood Communications, a business writing and editing company in Minneapolis.