Lessons learned while implementing artificial intelligence

Information given by tools should match what clinicians need, says UCI Health expert

IN THE NEWS: While artificial intelligence tools in electronic health records have helped physicians reduce some of their administrative work, Dr. Deepti Pandita, chief medical informatics and artificial intelligence officer and vice president of clinical informatics at UCI Health, tells Becker’s Hospital Review about some of the lessons learned.

“The patient-level clinical summary [generated by the EHR] is often not useful because, even when it reflects recentness, it is typically generated primarily from aggregated chart content rather than from the structured encounter-level decision outputs clinicians rely on: medication changes, newly placed orders and referrals, and the specific follow-up and return plan.”

While the AI is doing exactly what it was built to do, a diluted summary forces clinicians to reread the visit documentation to fill in the missing information, she says.

Pandita is a national leader in the field of clinical and medical informatics and currently drives artificial intelligence governance, digital transformation and clinician efficiency initiatives. In 2024, she co-authored a position paper by the American College of Physicians, clarifying the organization’s position on AI technologies and providing recommendations for their use when giving care.

Her digital solutions have improved the health of communities, eased the cognitive burden of clinicians and promoted the wellness of the clinical workforce. Pandita’s work has also reduced the average patient stay in the hospital, saved costs and improved operational efficiencies. She is one of Becker's Hospital Review’s Chief Medical Information Officers to Know. She was also named a Changemaker by the Health Information and Management Systems Society (HIMSS) in March 2026.