Wolf&Grayson EditorialHealthcare workers are drowning in paperwork. Between documentation requirements and administrative tasks, actual patient care keeps getting squeezed out. Dr. Samuel Hammerman has watched this problem get worse over 30 years in critical care medicine. As Chief Medical Officer at Select Medical, he is now working with Andor to fix it using AI.
Understanding What AI Is Fixing
Dr. Hammerman still works as a critical care doctor. That matters because he is not just theorizing about healthcare AI from an office. “We are discovering opportunities by leaps and bounds on a daily basis where AI can be utilized as a tool to gather information, bring it in, and then create actionable pathways around diagnosis, therapeutics potentially, and certainly workflows,” he explains.
The goal is straightforward: make life easier for the people actually taking care of patients. Most healthcare technology fails because it does not fit how doctors and nurses really work. Dr. Hammerman has seen enough bad implementations to know what not to do. Technology has to solve real problems, not create new ones.
Exploring Emerging AI Agents
Plenty of hospitals are nervous about AI agents, and that reaction is understandable. But Dr. Hammerman has seen how these tools can change things when they are designed correctly. “Andor is listening, right? They are paying attention both figuratively and physically and pragmatically,” he says. “They are able to draw in those conversations, take the data, and integrate it in a way that is meaningful.”
What comes out of all that listening is what he calls agentic AI. You can think of these agents as digital helpers that take over the repetitive work. A hospital with 100 staff members can suddenly feel like it has 150. Not because anyone is working harder, but because routine tasks are no longer consuming everyone’s time.
Ensuring Data Quality and Safety
Building AI for healthcare is nothing like building a shopping app. The margin for error is essentially zero. “Data drives this, it is a learning model. So all that information has to be refined and validated,” Dr. Hammerman explains. “Anything in the clinical space must be validated, reliable, safe, and reproducible.”
Medicine does not deal in maybes. “There are only two numbers in medicine: zero or 100. Something occurs 100% of the time, so you can reduce the risk to 0%. That is what we deal with,” he points out. Every AI agent that touches patient care has to meet that standard.
Getting there takes serious work. “You do not run a marathon by getting up out of bed one morning and saying, I am going to go 26 miles,” he says. “You do it step by step. And the finished product is not what you thought would be realized. It comes through that hard work of refining the information and making certain it is safe, reliable, and advantageous for the patient.”
Healthcare AI also has to clear strict regulatory and privacy bars. HIPAA compliance is not optional. Andor meets that requirement, handles multiple languages, and integrates with existing systems. No one wants to rip out an entire tech stack just to add one new tool.
The whole point is to give time back to clinical staff. When AI handles routine data gathering and basic communications, doctors and nurses can focus on the complex situations that actually require human judgment. It is not about replacing people. It is about letting them do the work they trained for instead of drowning in administrative tasks.
Connect with Dr. Samuel Hammerman on LinkedIn to follow more insights on practical healthcare AI.
