Payers nationwide are striving to operate more efficiently, while maintaining high levels of member and provider satisfaction. Success stems from reducing the friction and costs associated with core processes, like utilization management, appeals, medical necessity research and documentation for newly diagnosed patients.
Using AI solutions to augment the work done by humans is an attractive solution for many payers.
Getting started with these technologies, however, can feel daunting.
To learn more about what it takes for payers to successfully incorporate AI and support more members, Becker’s Healthcare recently spoke with Chris Caramanico, CEO of Elligint Health, Amy Qureshi, RN, executive vice president of product strategy at Elligint Health, and Steven Tolle, chairman of the board at Elligint Health. Mr Tolle has significant experience developing and implementing AI from his time at IBM, Merge and IgniteData and addition serves as Chief of AI Strategy at Elligint Health.
Data enables better care
When it comes to approving care and predicting member risk, AI tools can play a central role.
Mr. Caramanico explained that if an AI tool can tell a payer with 99% accuracy that it will approve care, it makes sense to implement systems and processes that speed decision-making. “There’s a patient and provider at the other end of the line,” he said. “Faster approvals drive Star ratings, as well as member and provider satisfaction.”
AI also has the power to predict disease progression and the likelihood of acute events. Those insights are critical for proactive interventions that reduce the cost of care. True predictive clinical analytics include intelligent workflows that respond dynamically to members at risk and drive the right next action.
Robust payer solutions deliver AI tools that can quickly summarize what’s known about a member, as well as recommend engines that enable AI agents to suggest the recommended next step.
AI systems alone, however, aren’t enough. Timely access to data is what truly makes it possible to transform payer operations with AI. Without that information, it’s impossible for approvers to understand the full member picture and identify medical best practices.
Many health plan policies include specific time frames for approvals. As a result, fast access to member data and insights is foundational for more effective utilization management.
“If plans don’t respond in a certain period of time, care must be approved by default,” Mr. Caramanico said. “Because payers don’t have all the data and the clock is ticking, they often end up approving care anyway.”
A human in the loop
To adopt AI safely, payers must embrace a human-in-the-loop approach. “There’s never a point where we want to fully automate machines making all the decisions,” Mr. Caramanico said. “That’s a step too far.”
Humans in the loop are more than manual workers in a computerized setting. Instead, payers must transform employees into knowledge workers who are assisted by AI. To realize this vision, organizations need to focus on the design, implementation and integration of AI tools into their workflows.
“Transparency is critical,” Ms. Qureshi said. “Processes need to be clear to people. When implementing AI, payers must think about how employees want assistance so the technology makes sense and will help them be more consistent when serving members.”
What leads to AI success
Before payers proceed to layer AI models on top of their existing processes, they must first examine the data sources and evaluate data integrity.
“In the old world of computers, we said ‘garbage in, garbage out.’ In the world of AI, it’s ‘garbage in, garbage gets magnified.’ The better the data source, the better the results will be. Getting your data house in order is the most important thing,” Mr. Caramanico said.
The next step is differentiating between high-level and low-level use cases. To deliver meaningful results, organizations must double down on the use cases that provide the greatest value. Focusing on early wins is also a best practice.
“Make sure the use case provides value and curate it so it hits the expected return,” Mr. Caramanico said. “At the same time, you must be careful not to sell the idea upstream that AI will change everything and take out 40% of the management costs. We aren’t there yet.”
Finally, before beginning any initiatives, payers need to establish an AI governance model. This is critical, given the hype around what AI can accomplish in utilization and claims management.
Many companies are building apps, for example, that can automate workflows, which include listening to conversations with members and suggesting documentation that can be added to the system of record.
“When you start using an AI tool with magnification capability, haste can be deadly,” Mr. Caramanico said. “Organizations end up getting burned or not getting the results they’re looking for.”
Building an ‘ecosystem’
Agentic AI technologies and orchestration platform-related standards, such as the Model Context Protocol (MCP), are still in the early stages. Mr. Tolle likened the current AI technology landscape to an arms race. He emphasized that in this environment, payers must avoid one-size-fits-all strategies and instead build flexibility into their infrastructure.
“You can’t lock yourself into one paradigm,” Mr. Tolle said. “If you aren’t flexible, you may miss out on opportunities. For instance, it might be better to use a small language model that’s heavily tuned for your use case, rather than a huge large language model.”
A better approach is to build an ecosystem that supports a variety of AI agents, ranging from those that organizations build in house, to those built by key vendors, including claims processing systems, clinical management systems and pharmacy management systems.
While it takes a village to build an ecosystem, payers must ensure that the content is heavily curated and reviewed. This means validating AI agent results to ensure they are accurate and can be trusted by users. Identifying the guardrails that technology partners have in place and how these guardrails factor user feedback into their models is also important.
Elligint Health publishes its AI Point of View for clients and prospects. In addition, the company has established strategic and client advisory boards to identify which payer use cases are high value and which are more commoditized.
When payers adopt an AI engine, organizational readiness is far more important than moving quickly. That means managing expectations about what AI can do and finding a trusted partner with proven experience.
“Payers must be careful to establish the right objectives, integration and guardrails,” Mr. Tolle cautioned. “That’s where Elligint Health is focused.”
To learn more about Elligint Health’s perspective on responsible AI adoption for payers, including governance, readiness and practical use cases, explore the company’s AI point of view.
At the Becker's 5th Annual Fall Payer Issues Roundtable, taking place November 2–3 in Chicago, payer executives and healthcare leaders will come together to discuss value-based care, regulatory changes, cost management strategies and innovations shaping the future of payer-provider collaboration. Apply for complimentary registration now.