Originally published in Becker's Behavioral Health.
For revenue cycle teams in behavioral health, eligibility issues are rarely standalone tasks. A single denial can trigger a series of manual processes: verifying coverage across payer portals, coordinating updates across multiple systems, and contacting members—all while working to avoid billing delays and maintain a positive member experience.
At Headspace, these workflows were both highly complex and operationally essential. As a result, they were also extremely difficult to automate—until the team discovered high-precision AI agents for behavioral health that were trained specifically on these processes.
Nancy Le, who leads pre-access and member billing, describes success in her role as balancing operational efficiency with member experience. Her team is responsible for ensuring coverage is accurately verified upfront and that members are billed correctly and promptly. When that process breaks down, the consequences affect both operational performance and the member experience.
“We want to submit the claim, have it process, and bill the member in a timely manner,” she said. “If claims keep denying, that delays everything—and we don’t want members getting billed months later for something they thought was already resolved.”
Before automation was introduced, termed eligibility workflows followed a familiar but labor-intensive process. If a member’s insurance had terminated and the issue was not identified beforehand, the claim would be denied or rejected. Nancy’s team would then manually verify eligibility, log into payer portals, confirm coverage status, contact the member for updated insurance information, and make updates across multiple internal systems.
This work was managed by a small team, often requiring navigation across several platforms to resolve a single case. “We have multiple systems where we store information,” Nancy explained. “You have to go here to pull one piece of data, then go somewhere else, and then go somewhere else again to email the member.”
The challenge extended beyond the number of systems involved. It also stemmed from the coordination required between them. Headspace relies on a mix of external payer portals and internal platforms. Certain actions must occur in a specific sequence, while payer-specific rules and contractual nuances frequently introduce exceptions.
“The order was very important,” Nancy said. “You had to do this step first before that step. And there were so many different scenarios—we had flowcharts with arrows going everywhere.”
To solve this challenge, Headspace partnered with Magical to automate the termed eligibility workflow. Rather than automating a single task, the goal was to replicate the entire process—from identifying terminated coverage to updating systems and initiating member communications.
Today, AI agents on the Magical platform manage much of that workflow end to end with more than 95% accuracy. The system verifies eligibility, conducts insurance discovery to identify alternative coverage, updates internal records, and contacts members when necessary. According to Nancy, this comprehensive approach set the platform apart from other solutions the team had considered.
“We had looked at solutions that could verify insurance or tell us the denial reason,” she said. “But we would still have to go in and do all the work ourselves afterward. This was the first time we saw something that could actually handle the full process from start to finish.”
Nancy also described her initial skepticism toward AI, which stemmed from uncertainty about how it would perform within detailed, real-world workflows. What ultimately changed her perspective was the transparency of the system and the ability to make rapid adjustments.
“Seeing the test runs step by step really helped,” she said. “You can see exactly what it’s doing at each point. And when we asked for changes—even very small ones—they were implemented quickly. That made it easier to trust the process.”
In some cases, the system has uncovered opportunities that would have been difficult to identify manually. Because insurance discovery is built into the workflow, it has already located alternative coverage that prevented unnecessary outreach and billing.
“There was a scenario where it found new insurance for us,” Nancy said. “That saved an outreach, avoided a confusing bill for the member, and saved work for our team.”
Although volumes for termed eligibility have initially been lower than expected due to improvements in denial prevention upstream, the accuracy and transparency of the Magical platform have encouraged Headspace to explore automation across additional workflows, particularly those involving member communications.
“We’re excited to expand,” Nancy said. “It’s essentially building on what’s already working for us today.”
For organizations considering similar initiatives, Nancy stresses the importance of keeping an open mind and taking advantage of structured testing before deployment.
“Don’t feel overwhelmed by it,” she said. “There are ways to test, validate, and make sure it’s doing what you want before you fully roll it out. Once you see it in action, it changes how you think about what’s possible.”
As healthcare organizations continue to operate in increasingly complex environments, coordinating workflows across systems—and doing so consistently—remains an ongoing challenge. For Headspace, the transition has been about more than reducing manual effort. It has been about creating a more reliable and scalable approach to executing critical processes that directly affect both revenue and the member experience.
To learn more about how Magical applies agentic AI to automate administrative workflows across healthcare operations, visit getmagical.com.