Tech leaders’
AI mandate,
do more with less

Summary

• Members no longer compare your experience to other health plans—they compare it to the quick answers they get from generic AI chatbots.
• Getting “unstuck” comes down to three things: aligned leadership, a business case that holds up, and the right partner.
• On build-versus-buy, the deciding question isn’t “can we build it.” It’s what gets members to a safe, healthcare-grade experience the fastest.
• Success is when members choose your experience on their own and keep coming back.

In payer organizations across the U.S., technology leaders are facing what feels like an impossible mandate: ship an AI-driven member experience, with the same budget and headcount as last year.

The pressure isn’t just coming from the boardroom. It’s coming from members who already ask ChatGPT about a new symptom, a medication interaction, or a bill they don’t understand, and the answers come back instant and convincing. That’s the new bar. Members aren’t measuring you against your competitors anymore, they’re measuring you against the best answer they got last night from a tool that knows nothing about their benefits or their clinical history. When that answer beats yours, they stop coming to you for guidance.

At a recent roundtable, League’s CTO, Dan Galperin, sat down with Raj Kurup, VP of Operations & Engineering at Point32Health, alongside technology leaders from payer organizations across the U.S. to talk about how they’re approaching this mandate. The conversation was candid about why so many organizations get stuck, what it takes to move faster, and how to build the experience members expect.

Here’s what came out of that conversation.

Getting “unstuck” comes down to three things

Most plans aren’t stalled because the technology isn’t there. They’re stalled on legacy infrastructure, a lack of internal alignment, and the scars of past transformation efforts that didn’t stick. 

Raj has firsthand experience navigating this mandate and defining a path forward. What breaks the pattern, from his perspective, is a “perfect storm” of three things:

• A convergence of leaders who genuinely care about the member experience and are ready to do the hard work. 
• A business case where the math holds up, and is clear enough to move on without getting lost in the details.
• The right partner: a team of trusted experts that accelerates time-to-value and is committed for the long haul.

Without these, the work becomes another transformation project that doesn’t stick.

There’s more to build-versus-buy than “Can we build it?”

Many technology leaders have an instinct to build. But when they run the build-versus-buy assessment on an AI-driven member experience, owning every piece isn’t realistic.

Modern AI tools make it relatively easy to build impressive-looking prototypes. The hard part is the plumbing underneath: certifications, clinical guardrails, PHI-aware access, escalation rules, and more that keep member-facing AI safe. 

Dan explained that on League’s agentic solutions, the majority of engineering time goes not to the conversational layer but to guardrails, compliance, security, and answer quality that ensures zero hallucinations. That’s the iceberg beneath a “simple” AI feature – one that most health plans don’t have the resources to build and then maintain, pulling your best people away from the most important work that actually moves the needle on the member experience.

And every hour spent there is an hour not spent innovating. So the deciding question isn’t “can we build it.” Nearly everyone can build something. It’s what gets an organization to a safe, member-facing AI experience the fastest, and leaves room for innovation.

Member-facing AI needs one unified front door   

Getting build-versus-buy right is only part of the challenge. The other is ensuring a unified, seamless experience. Over the years, most payers have accumulated a long list of point solutions that didn’t share data, weren’t easy for members to find, and didn’t make a measurable difference. The same pattern can repeat with AI. When efforts launch in silos, payers risk ending up with dozens of disconnected tools, rising token costs, and a member experience that still feels fragmented.

Raj was direct about the consequence: when AI efforts are scattered, members see the disorganization, and that’s the opposite of the experience plans are trying to build. 

Routing everything through a single platform solves for that. Members get one consistent experience on the surface, and because the platform is open, it integrates with the systems you already have and lets you add new AI capabilities over time, instead of forcing a rip-and-replace. That is what keeps the experience unified as it grows, and stops fragmentation from creeping back in.

The complexity is real, but it’s no longer an excuse

Members are already taking their health questions to general AI chatbots, with three in five U.S. adults saying they’ve used AI tools for health questions in the past three months. The answers come back friendly and confident, but they’re often wrong, incomplete, and not personalized. 

Health plans have an advantage here. As Raj pointed out, a tool like ChatGPT isn’t a real health advisor, because it isn’t contextualized. It doesn’t know a member’s coverage, benefits, or care history. A plan knows all of it. That context is what lets plans deliver guidance a member can actually trust and act on.

The trap is treating the complexity of healthcare as a reason to move slowly—or not at all. But complexity no longer justifies year-long decision cycles or experiences that frustrate members in the moment they need help most. Our favorite line from Raj: “Don’t fall in love with the problem. Fall in love with the solution.” It’s easy to get absorbed in how hard the problem is. The key is staying focused on what you’re going to do about it.

When thinking about success, put yourself in your members’ shoes

Ask most organizations what success looks like, and you’ll get a set of metrics: adoption rates, call deflection, engagement. Those matter, but Raj offered a simpler and more demanding test. Would he use it himself?

The real measure of success is whether people choose the experience on their own, and come back to it because it helped. Members have easy alternatives for their health questions, and they’ll use whichever one serves them best in the moment. Building something people return to without being prompted is the hard part, but it’s the kind of success that lasts. 

The mandate is hard, but it isn’t impossible

The plans making real progress have a few things: a leadership team that is truly aligned, a bias toward speed and value, and a steady focus on building an experience that is unified, safe, and genuinely helpful.

This is also the thinking behind League’s self-serve console, which lets teams define what an agent does, connect it to the data it’s permitted to use, set guardrails and escalation rules, and validate it against synthetic populations, all without writing code. Because the healthcare-grade foundation is already built into League’s platform, every agent starts safe. Safety and speed stop being a trade-off, freeing teams to focus on building an experience that sets their plan apart.

Because behind every stalled transformation are real people waiting for better care: a member trying to understand a diagnosis, access their benefits, or take the next best step in their healthcare journey. Keep them at the center, and the hard decisions get clearer.