Healthcare has been chasing digital transformation for decades, and frankly, most of it has been disappointment layered on top of legacy systems. Electronic health records migrated patient data in the early 2000s but left it fragmented across siloed departments. Telehealth removed geographical barriers but never replicated the quality of in-person care. Now a new wave is cresting: according to KPMG, 68% of healthcare providers have already adopted AI agents into their workforce—and the results from early adopters are starting to speak for themselves.

The Claims Processing Win That Started It All

Hospital for Special Surgery (HSS), an academic medical center in New York City specializing in musculoskeletal care, provides one of the most concrete case studies available. Their AI agents now handle 1,100 insurance claims per month—work that previously took several weeks and required both internal staff and a third-party contractor to manage volume. The numbers are striking: appeals processing dropped from 45 minutes to just five, and the success rate climbed from 65% to an impossible-sounding 100% in the nine months since implementation. HSS now handles all claims entirely in-house. "Agentic AI takes your workflow and collapses it, augments it, supercharges it, and makes it more performant," says Dr. Ashis Barad, chief digital and technology officer at HSS.

From Back Office to Patient-Facing: The Scheduling Triage Deploy

Building on that success, HSS partnered with enterprise agentic AI developer Ema Unlimited to deploy a patient-facing scheduling and triage service accessible 24/7 via web, text, or phone. The system uses conversational AI to ask clarifying questions about a patient's condition before booking appointments with the most appropriate clinician—factoring in location, insurance coverage, and physician availability. Dr. Barad describes it as completing "the whole loop." The agent is trained on all HSS context, rules, and knowledge base, giving patients streamlined access to specialist expertise from world-leading surgeons without manual routing overhead.

Guardrails for High-Stakes Decisions

Given the stakes involved in healthcare decisions, HSS built in safeguards. Sensitive, complex, or uncertain scenarios automatically escalate to human specialists. Every AI decision remains auditable, and staff can intervene at any point. Patient data stays secure, and the system trains on all HSS protocols, policies, and care pathways. Dr. Barad emphasizes that patient-care-adjacent agents face far more scrutiny than backend processes—decisions flow through an AI subcommittee he co-chairs alongside a senior nursing executive. "It’s wrong to think of agentic AI in use cases," he argues. "It’s a general-purpose technology, analogous to electricity."

The Data Foundation Problem Nobody Talks About

That analogy only works if you have the infrastructure to support it—and that's where most healthcare organizations are still falling short. HSS's patient-facing agents can draw from clinical care history and existing clinician recommendations, combine that with current symptoms, and decide whether escalation is needed before notifying specialists—all because they've invested in data interoperability. Dr. Barad points out that even basic metrics like "time to start surgery" have varied definitions across hospitals he's worked in. Fragmented data sources with nonstandardized definitions prevent AI agents from retrieving information or building the tacit knowledge base that differentiates them from simpler automation tools.

The 90% Vision and Why It Matters

Dr. Barad's vision is bold: he estimates 90% of non-clinical healthcare tasks could eventually run through AI agents, freeing clinicians for what he calls "white-glove work"—the most complex, specialized, and sensitive cases. KPMG research suggests the industry agrees; 84% of providers say they're comfortable delegating specific process decisions to AI agents already. "We're spending so much time on keyboards and computers right now that we're actually not doing what we should be doing," Dr. Barad says. "This is going to rehumanize healthcare."

Key Takeaways

  • KPMG data shows 68% of healthcare providers have adopted AI agents, with 84% comfortable delegating decisions to them
  • HSS reduced claims appeals from 45 minutes to 5 and hit 100% success rate using agentic AI for insurance processing
  • Patient-facing triage services now handle scheduling 24/7 via web, text, or phone at HSS through an Ema Unlimited partnership
  • Healthcare leaders emphasize treating agentic AI as general-purpose infrastructure, not isolated use cases

The Bottom Line

The healthcare industry has been burned by overpromised technology before—but these numbers from HSS are hard to dismiss. When you can point to specific metrics like 45-to-5 minute improvements and 100% success rates, you're past the hype cycle. The real challenge isn't whether agentic AI works; it's whether hospitals will do the unglamorous work of fixing their data foundations first. Those that do might actually deliver on the rehumanization promise.