Primary care in the United States is facing more problems. Almost 100 million Americans do not have a regular primary care doctor. Because of this, about 2 billion hours of primary care go unmet every year. People cannot fill this gap by working harder alone. Doctors have too many patients, lots of paperwork, and complicated schedules. They spend a lot of time on tasks that are not direct patient care.
At the same time, hospital and clinic leaders try to make workflows smoother and lower costs. Many staff members are responsible for scheduling appointments, checking medications, contacting patients, and helping after hospital stays. They do not always have the right resources. This causes doctors to get tired, care to be delayed, and sometimes worse results for patients.
These problems show a clear need for ways to make work easier, keep doctors available, and keep patients involved. Artificial intelligence (AI) can help when it is added properly into current clinical workflows.
AI that works inside regular clinical routines can do more than just analyze data. It can manage patients in real-time. Unlike old systems that only report facts, embedded AI can take immediate steps that help patients and make doctors more efficient.
For example, Tom™ by Lumeris is an AI solution supported by over 20 years of research and 5 years of AI work. It works all day and night with healthcare teams. Tom can set up screenings and appointments, check if patients take their medicine, reach out after hospital stays, and give patient education. When cases are hard, Tom alerts human doctors.
Tom uses information about patient history, clinical rules, and social factors to teach and act in ways that fit each patient’s needs. This helps care teams handle more cases and lets doctors spend time on patients who need their attention the most.
John Doerr, chair of Kleiner Perkins, says Tom can help reach the 100 million Americans without primary care and cut healthcare costs by more than half. It works without changing how doctors normally do their work because it is part of the existing system.
By working inside electronic health records (EHR) and clinic software, AI like Tom saves time on office work and lowers mistakes in scheduling. This makes doctors happier because they spend less time coordinating and following up.
One important way embedded AI helps is by managing schedules. The number of patients a doctor can see depends on the doctor’s availability. Embedded AI helps scheduling by:
These features make the clinic run better and let doctors do more work. Mike Long, CEO of Lumeris, says Tom helps fix the 2-billion-hour primary care shortage by organizing tasks and allowing more effective care time.
AI also helps automate more than just scheduling and contacting patients. It helps with clinical and office workflows in many ways, such as:
Tools that understand language now help write clinical notes, transcribe speech, and code medical information. Microsoft’s Dragon Copilot is one example. This AI helps create referral letters, after-visit summaries, and clinical notes based on evidence. Automating these tasks saves doctors time and lets them focus more on patients.
AI can check and code insurance claims faster and more accurately than people. This reduces errors, claim denials, and payment delays that slow down clinic money flow. Hospitals and clinics save millions by making billing faster and reducing mistakes.
AI looks at patient health and payment data to predict health risks or billing problems. This helps doctors act early, which can improve patient health and money collection. For example, AI finds patients who might miss appointments or have trouble taking medicine so care teams can reach out to them.
AI customizes how it talks to patients based on their needs, preferences, and social factors. This helps patients follow treatments better and promotes healthier habits. Community health programs in low-income areas especially benefit by using AI to focus on social and economic factors.
The United States has many groups working to add AI in health care. Lumeris, started in 2012, has over 1,200 engineers, doctors, and health specialists. Their AI tool Tom shows clear improvements in primary care. Lumeris’s Medicare Advantage plan, Essence Healthcare, has earned strong CMS ratings (4.5 to 5 stars) by using AI to improve quality and patient experience.
Health providers, schools, and tech companies work together to build AI faster. For example:
These partnerships show that AI workflows are becoming common in clinics, especially in big, integrated health systems that care for complex patient groups.
Recent studies show AI use in healthcare is growing fast. A 2025 AMA report says 66% of U.S. doctors use AI tools. This is up from 38% in 2023. Also, 68% of those doctors think AI helps patient care. This growth includes both tools doctors use directly and office automation behind the scenes.
Healthcare AI is expected to grow from $11 billion in 2021 to nearly $187 billion by 2030. This happens because more people want systems that help with diagnosis, patient contact, smooth clinic work, and billing.
These changes match the needs of clinic leaders and IT managers who want to reduce work complexity while increasing doctor productivity and patient satisfaction.
Even though AI has many benefits, there are challenges to adding it. It can be hard to connect AI with current Electronic Health Records (EHR) and clinic systems. It may take big investments and changes. Standalone AI tools can interrupt work instead of helping. That is why AI should be built right into clinical operations.
There are also concerns about ethics and rules like protecting patient privacy, avoiding bias in data, and who is responsible when things go wrong. Groups like the U.S. Food and Drug Administration (FDA) make rules to keep AI tools safe, clear, and fair. Clinics using AI must follow these rules to stay legal and ethical.
These uses reduce paperwork and improve care quality, which helps clinic managers handle many tasks better.
Adding AI into clinical workflows offers a clear way forward for healthcare in the United States. By automating patient tasks and improving schedules, AI helps doctors and clinics meet more demand and lower costs. As healthcare, technology, and research teams keep working together, AI will become a bigger part of primary care and medical practice management.
Tom is an AI-powered Primary Care as a Service™ solution developed by Lumeris to support overburdened physicians and expand primary care access by integrating into clinical workflows and executing autonomous patient management actions.
Tom tackles primary care provider shortages, administrative burdens, limited patient access, and the growing mismatch between demand and supply by expanding care capacity and proactively managing patient care tasks.
Unlike traditional systems that only provide information, Tom autonomously acts on data in real-time, scheduling appointments, monitoring medication, conducting outreach, and personalizing care within shared care plans.
Tom autonomously schedules screenings and appointments, manages care coordination, monitors ongoing patient needs, and escalates complex cases, effectively optimizing provider schedules and reducing administrative workload.
Tom incorporates social determinants of health data alongside clinical guidelines to personalize patient interventions, improving engagement and outcomes while addressing non-clinical factors impacting health.
Tom aims to bridge the 2-billion-hour annual shortage in primary care by expanding provider capacity and enabling access for an estimated 100 million Americans without primary care providers.
Collaborators include BJC Health System’s Center for Health AI, Endeavor Health, MIT Computer Science and AI Lab, Oliver Wyman, ŌURA, and Wolters Kluwer, bringing expertise in AI, healthcare delivery, decision support, and wearable tech.
Lumeris reports improved quality metrics, better patient experiences, enhanced physician satisfaction, and high CMS star ratings (4.5 to 5.0) across multiple Medicare populations using Tom.
By increasing primary care capacity, reducing administrative burdens, and enabling proactive patient management, Tom lowers care costs by over 50% through improved efficiency and prevention.
Being embedded allows Tom to operate 24/7 alongside care teams, providing real-time insights and taking immediate, appropriate actions without disrupting provider workflows, thus enhancing schedule management and care delivery.