June 2, 2026

After XGM 2026: What Health System Leaders Need to Know About Agentic AI Before Implementation Begins

XGM is always a great time for health system leaders, operational teams, and technology experts to come to Verona and get energized about what Epic is building next. The harder reality begins once organizations return home and start evaluating what it will take to operationalize those ideas inside complex healthcare environments. That transition from conference excitement to operational execution felt especially pronounced this year as agentic AI dominated nearly every conversation. 

While many discussions focused on what AI agents can do, the more important questions centered on governance, trust, economics, and implementation at scale. Here are the four key themes we heard from customers and prospects stopping by the booth. 

1. Agentic AI will encourage organizations to keep as much self-pay AR as possible 

For years, health systems have outsourced portions of self-pay AR because pursuing smaller balances internally often wasn’t worth the labor required. Agentic AI is beginning to disrupt that equation. 

Epic already sits on the core account data, and with Hello World and Cheers expanding AI-driven outreach capabilities, organizations are starting to evaluate whether portions of self-pay recovery can be brought back in-house using AI agents instead of traditional outsourced models. In the year ahead, more health systems will compare the cost of AI-driven workflows they already own against the rates they pay organizations built primarily around offshore labor. 

That shift creates significant pricing pressure for vendors trying to remain relevant. The vendors that survive this shift will be the ones bringing something agents simply cannot replace. The market is moving from labor-based services toward intelligence-based services, where long-term value will come from helping healthcare organizations operationalize AI safely within highly regulated financial workflows. 

That includes compliant payment infrastructure, trusted patient communications, provider and payment intelligence, and workflow governance. In this environment, the partners that remain strategically valuable will be the ones that make AI more effective, trustworthy, and operationally sustainable.  

2. The cost structure of agentic AI is still an open question

Adoption is accelerating, but the economic and governance models underneath it are still taking shape. As with the broader AI market, healthcare leaders are already asking important operational questions: 

• What tasks should agents be allowed to perform, and how narrowly should those workflows be constrained?  
• Who governs escalation paths and maintains accountability when something goes wrong?  
• How will organizations monitor usage and costs before they spiral at scale?  

The reality is that healthcare organizations are unlikely to unleash fully autonomous agents across sensitive financial and patient workflows anytime soon. Directed, highly governed use cases will define the near-term adoption curve as organizations work to balance innovation with operational, financial, and compliance risk. 

3. Trust and governance will define adoption

High-profile stories about AI agents behaving unpredictably have created legitimate caution across healthcare teams. Mistakes in healthcare workflows carry real operational and compliance consequences, particularly when patient balances, scheduling, or regulated communications are involved.  

As a result, healthcare organizations will keep humans in the loop and lean heavily on trusted partners to provide the infrastructure, expertise, and governance needed to ensure AI-driven workflows remain accurate, compliant, and sustainable at scale. The question every organization needs to answer now is not just which agents to deploy, but which partners are equipped to help implement them responsibly. 

4. Content and trusted data will continue to be king 

As AI agents integrate more directly into Chronicles, healthcare organizations need to answer two questions before scaling deployment: Is the underlying data trustworthy enough for agents to act on, and is there specialized intelligence outside the EHR that would improve outcomes? 

This is where MCPs and specialized content providers come into play. As AI models become more widely accessible, what will matter most is the quality of the data and governance supporting those models. 

Organizations will increasingly look for partners that can provide specialized payment intelligence, trusted provider data, compliance-aware workflows, and structured operational insights that improve agent decision-making at scale. 

The bottom line 

XGM surfaces what is technically possible. The harder work begins when leaders return home and start deciding which partners are equipped to help them get there. The healthcare organizations that choose partners that bring genuine compliance depth and specialized intelligence will be the ones that realize the full potential of agentic AI. 

As organizations evaluate the next phase of AI adoption, the conversation is becoming less about who can automate the most tasks and more about who can ensure those automations are accurate, governed, compliant, and operationally sustainable at scale. 

Curious how specialized payment intelligence, compliant engagement workflows, and trusted provider data can help make AI agents more effective inside Epic? Let’s connect.