This is why emerging standards around AI transparency, provenance and auditability matter. They lay the groundwork for a future where both clinicians and patients can understand when AI influenced care and how its limitations are managed.
Remote Patient Monitoring: A Stress Test for AI and Care Models
If you want to see where AI succeeds, or fails, look at remote patient monitoring( RPM). RPM promises continuous oversight for chronic conditions, post-discharge recovery and high-risk patients. At our health system, RPM programs have expanded across heart failure and hypertension and have been very successful in improving care in patients with those problems.
RPM can generate a constant stream of signals. Without prioritization, clinicians can drown in alerts. Without clear ownership, alarms go unanswered. Without workflow integration, data sits unused: this helps no one.
This is where AI can become essential. AI can triage signals, suppress noise, surface trends and route actionable information to the right role at the right time, but only with guardrails. RPM is a perfect example of why agentic AI must operate within explicit protocols, escalation paths and human oversight.
RPM also exposes a deeper truth: you cannot deliver continuous care using episodic workflows. We must have awareness of what goes on outside of our walls to best care for our patients.
How AI Is Changing the Provider Experience
For clinicians, AI succeeds or fails based on one question: Does this make my day better? When AI reduces inbox volume, prepares visit summaries, drafts notes and handles routine administrative work, it feels like a force multiplier. When it adds alerts, screens or uncertainty, it becomes another source of burnout.
At our organization, the most successful deployments are those designed workflow-first. AI isn’ t bolted on; it’ s embedded where work already happens. It’ s generally the only way to have a successful implementation. That’ s appropriate. AI now touches clinical judgment, not just documentation.
The Defining Theme of the Next Two Years: Speed vs. Absorption breakthrough product. They’ ll be defined by stacked improvements: documentation plus summarization plus routing plus task automation.
Organizations that thrive will be those that can adapt quickly without sacrificing trust.
That means investing in governance, transparency, monitoring and change management: not as stop signs, but as enablers of safe speed.
Where the Impact Will Be Greatest I believe we’ ll see the most immediate and measurable impacts in:
· Patient access and front-door operations: Scheduling, triage, messaging and referral coordination.
· Primary care: Inbox management, visit prep, documentation and chronic disease oversight.
· Imaging-intensive service lines: Emergency care, oncology pathways, stroke response.
· Inpatient transitions of care: Admission prioritization, discharge planning, follow-up coordination.
· Revenue cycle and utilization management: Prior authorizations, documentation support, denial response.
These are high-volume, high-friction areas where small efficiency gains compound rapidly.
The Bottom Line
AI in health care is no longer hypothetical. It’ s already shaping care— quietly, unevenly and increasingly. The real transformation isn’ t about machines becoming smarter. It’ s about health care becoming more capable of using AI responsibly.
The next two years will reward organizations( and their patients) that pair innovation with accountability, automation with restraint and speed with trust.
The quiet revolution is already underway. The question isn’ t whether AI will change health care. It’ s whether we’ ll change health care enough to use it well.
Dr. Oliver is the Chief Medical Information Officer for Baptist Healthcare System.( non-member)
AI capability is compounding. Health care’ s ability to absorb change is not. Health care is not known for its rapid evolution and historically this often serves patients well. That may no longer be the case. I believe the next 12 – 24 months won’ t be defined by a single
April 2026 11