AI in Oncology Practice: Innovation in the Delivery of Care Itself

For attendees at the 2026 Community Oncology Alliance (COA) Community Oncology Conference in April, artificial intelligence (AI) embodied the “Innovation in Practice” theme from start to finish, with mentions from the opening session to the closing minutes and everywhere in between.

Tapping AI’s potential has been a priority for COA President Debra Patt, MD, PhD, MBA, since she stepped into the role in January 2025. As she told The American Journal of Managed Care at that time, AI could ease some of the biggest burdens in oncology care, from staffing shortages to diagnostic delays. Realizing that potential, she said, would require true partnerships between AI creators and the physicians, nurses, and practice leaders who use technology. And this year, the COA conference doubled as a showcase for these new tools.

Patt, who has piloted and launched many technology solutions as executive vice president at Texas Oncology, continued that theme of partnership at the second of COA’s innovation showcases. “I see innovation and collaboration,” she said. “We recognize at COA that practices are really going to be challenged with how we have digital transformation, and so that's really the purpose of that being the theme of this conference. COA has always done a great job of meeting practices where they are and helping support us as we evolve.”

AI in Today’s Practice Environment

To start, however, COA attendees heard how physicians are already using AI in practice today. The conference opened April 28, 2026, with the session, "AI in Practice: Operational and Clinical Wins in Community Oncology,” moderated by Jeff Hunnicutt, CEO of Highlands Oncology, and featuring Larry Bilbrey, senior director of digital innovation for OneOncology and director of data insights and innovation for Tennessee Oncology, and oncologist David Oubre, MD, the founder of Pontchartrain Cancer Center (PCC) in Covington, Louisiana.

Hunnicutt opened with a level-setting reminder that AI is not new—it ranges from 1980s pattern recognition to modern generative models—but the pace of advancement has accelerated. He noted that the panel included a range of practice scales: PCC is a small independent practice, Highlands is a midsized group, whereas Tennessee Oncology is a large multisite organization.

“What we've seen is that it really starts [with] the philosophy,” Hunnicutt said. “You have to be able to understand why it is we're doing what we're doing. What are the guardrails that we're going to put around?”

Bilbrey said his approach in evaluating an AI tool is straightforward: Does it solve a real problem? He called for a “provider first” philosophy, arguing that improving physician workflow reduces burnout and ultimately improves patient care. Oubre described his approach to technology and AI adoption as incremental, preferring small, carefully managed steps that fit the specific needs of his practice.

Both panelists pointed to the use of ambient scribes as an early AI win. Oubre found that using AI to assist with documentation restored his ability to fully connect with patients, rather than looking at a computer screen.

“It's really about the physician-patient interaction—what’s going to work for that interaction, which is the core of what we do,” he said. Like so many physicians, Oubre said that the advent of the electronic health record had disrupted this element, as it required typing notes into a computer during the visit “with my back turned to a patient, because I can't type without looking…at the screen.”

With the ambient scribe, the recording replaces the typing. “And so that immediate interaction, almost the first time I did the scribing system, was ‘Wow, this is freeing.’ How great this is: I can sit here and talk to the patient, look them in the eye, and I felt better. I think the patients felt better, too, but it made me feel like I was really communicating with them.”

Bilbrey shared how Tennessee Oncology is taking on the technology dinosaur of the medical practice: the fax machine, which persists due to concerns about security with transferring certain records via email. With automated fax processing using AI, Tennessee Oncology is reducing handling time from roughly 10 minutes to under 2 minutes per document.

Oubre said AI scribes do require note editing, and the panel addressed the need for overall change management. Hunnicutt described how Highlands Oncology built a cross-functional technology committee to cultivate a culture of openness to change before introducing new tools.

Bilbrey warned practices to protect intellectual property in vendor contracts and avoid overly long agreements. “Technology prices just keep falling. So don't lock yourself into this massive contract that is going to cost you $200,000 a year, and then it's only going to cost them $10,000 to operate,” he said.

Oubre pushed back on fears that AI could replace physicians. “There's a talent that you get when you walk into a room, and if you've been doing this for years, you can tell when someone has something really wrong with them,” he said. “You can just see it right away. You know that patient, you know how they're doing,” The way 2 different patients say they are doing “really bad” will mean different things, and a physician who knows the respective histories will grasp those nuances. AI cannot do that, Oubre said.

“I think it will probably replace a few jobs, but I don't see the physician, the teacher, or people like that being replaced,” he said.

AI Improves Patient Experience—Starting With the Phones

At the COA showcase, Lucio Gordan, MD, president and managing physician of Florida Cancer Specialists & Research Institute, said the session offered an opportunity to absorb conceptual change that is no longer optional. Comparing AI to adopting the internet or a smartphone, he introduced the presenters, Canopy and Reimagine Care, around a single pressing need in community oncology: extending high-quality care beyond the office walls without adding unsustainable burden to already-stretched physicians, nurses, and staff.

Meet Kathy from Canopy. Andrew Frank, who leads the provider product team at Canopy, presented the company's work managing patient communication between visits—remote symptom monitoring, secure messaging, and inbound call management. Frank introduced Kathy, an AI-powered virtual phone operator currently handling calls for Highlands Oncology.

Kathy is activated during peak overflow moments, notably Monday morning rushes or lunch hours when staff step away. Kathy fields calls that would otherwise go to voicemail or be abandoned entirely. In a cheery voice, Kathy greets callers, verifies patient identity, identifies the purpose of the call, asks appropriate follow-up questions, and generates a structured ticket in Canopy's system for downstream staff. The COA audience listened as Frank played recordings of 2 real patient calls from Highlands (with patient consent and names redacted). The first involved a medication refill question, and the other was a request to reschedule an appointment.

Then, Frank demonstrated a live call with Kathy, taking a call in which the patient mentioned vomiting; this triggered Kathy's clinical guardrails and an immediate handoff to a live nurse, a feature developed in collaboration with the Highlands team.

Data from a recent busy Monday showed that Kathy handled nearly 20% of all incoming calls, which Frank said matched the capacity of almost 3 full-time employees flexing in on demand.

“We're excited as we look at what's next for Kathy. We think we're going to scale Kathy in terms of taking more and more phone calls, going beyond the overflow use cases,” Frank said. “There are opportunities to put Kathy at other places in the phone tree.”

Reimagine Care and the AI care layer. Dan Nardi, MS, CEO of Reimagine Care, said that for decades, innovation in oncology meant new therapies. Today, he said, “We have an opportunity to bring innovation to the delivery of cancer care.”

Nardi opened with headline outcomes drawn from tens of thousands of patients across community practices: a 70% reduction in emergency department visits, a 42% reduction in clinical inbox messages, 9-out-of-10 patient satisfaction, and a provider Net Promoter Score of 98—achieved without adding staff, increasing provider burden, or surrendering control of the patient relationship.

The mechanism is Remi, an agentic AI virtual assistant that operates as a persistent "care layer" between the patient and the practice, communicating entirely via SMS, with no app download or portal login required. Remi does 3 things simultaneously, as follows:

• Offers continuous patient engagement, which means proactively checking in and responding to patients within seconds;
• Conducts AI-augmented triage, which contextualizes patient messages against the chart and routes them to the appropriate level of care; and
• Connects the patient to human clinical oversight with oncology-trained medical assistants, registered nurses, and advanced practice nurses (APNs), prompting conversations within minutes when escalation is required.

Nardi illustrated this through a patient scenario: Jane is a 73-year-old woman with triple-negative breast cancer. On day 4 of her first cycle of dose-dense AC (doxorubicin plus cyclophosphamide ), she notices redness and warmth at her incision site. It is a Thursday evening, and the nearest emergency department is 45 minutes away. At 8:35 PM, she texts Remi, and the interaction is fully triaged through SMS in 6 minutes, with zero minutes of human time required at that stage. By 8:41 PM, a nurse takes over; by 8:50 PM, an APN has conducted a video visit; by 9 PM, a prescription for clindamycin has been sent to a local pharmacy. Remi follows up automatically on Friday morning and Saturday to confirm improvement.

Reimagine Care emphasizes that its solutions are designed to work within existing systems and with current employees, not as replacements. The presentation ended with the question that abounds about these tools: “Will AI remain a tool…or become part of the care team?”