Philadelphia-based Penn Medicine has budgeted for $105 million in benefits from AI solutions by fiscal 2028.
The technology is already delivering value at the academic health system in areas like radiation oncology treatment planning and clinical documentation, IT leaders told Becker’s.
“We generally do not do [AI projects] unless there’s a clear benefit — a patient safety benefit, or a dollar benefit on the revenue or operations side,” said Srinivas Sridhara, PhD, chief data and analytics officer of Penn Medicine.
One measurable success has been an internally developed tool that automates contouring to target radiation therapy, a job typically done by a dosimetrist.
“What you might have needed an expert clinical teammate to do is now being done fully in the workflow with a machine learning tool,” Dr. Sridhara said. “There’s an increasing demand for radiation therapy, but simply not enough dosimetrists out there.”
The seven-hospital, $12 billion system is implementing AI for EHR search, prior authorizations and quality registries, with mature ambient AI and imaging tools in production. “This is not a theoretical, flash-in-the-pan thing — it is really useful and practical,” Dr. Sridhara said.
Penn Medicine frames the technology internally as “innovation with guardrails,” continuously monitoring bias, performance and value, he said. “A lot of organizations use governance to get to ‘no.’ We’ve been trying to get to ‘yes,’ while managing risk.”
The organization has more AI use cases in the pipeline, but the ROI figures include what it’s formally committed to.
“If you take something like guideline-directed therapy for heart failure — very under implemented, very clearly tied to improved outcomes — and if you can bring those closer to the point of care in a more accessible, consumable fashion for our care teams, there’s a lot of opportunity there,” said Sri Adusumalli, MD, vice president and chief health information officer of Penn Medicine. “We’re looking for these tools to produce outcomes, and being able to tie the tool to those outcomes is something we’re starting to do.”
For instance, Penn Medicine has seen as much as a 20% productivity improvement among project managers using AI copilot tools to handle coordination, documentation, and cross-team work, according to Dr. Sridhara.
Going forward, Dr. Adusumalli envisions AI being woven into each stage of care, from triaging patients to virtual or in-person appointments, synthesizing previsit information for providers, and guiding their treatment and diagnostic decisions.
“Having a solution that ties the patient and provider journey together — rather than independent pieces — would be amazing,” he said.
Added Dr. Sridhara: “If neither money nor regulation were an object … we could build a way to fully deliver basic care with AI agents.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
