March 28, 2026
By Jay Chowdappa, MD, co-founder and CEO, Zynix AI ยท Updated October 1, 2026
To automate care management with AI agents, hand the repeatable work to agents (outreach calls, scheduling, reminders, routine check-ins, referral tracking and documentation preparation) and keep clinical assessment, care planning, crisis response and escalation decisions with care managers. Start with one standardized, high-volume workflow such as TCM, measure it against a baseline, then expand.
Care managers carry large panels but spend much of their time on administrative tasks: making outreach calls, scheduling appointments, documenting interactions, tracking referrals, and chasing down information across systems. That leaves too little time for the complex clinical judgment that actually requires human expertise.
Automate: Initial outreach calls, appointment scheduling, medication reminders, barrier screening, transportation coordination, routine follow-up calls, documentation of standard interactions, and referral tracking.
Keep human: Complex clinical assessment, motivational interviewing for behavior change, crisis intervention, care plan design for multi-morbid patients, family/caregiver coordination for complex cases, and clinical escalation decisions.
AI agents automate the repeatable parts of the 30-day post-discharge TCM workflow: early outreach after discharge, follow-up appointment scheduling, 7-day and 14-day check-ins, and barrier resolution. Licensed clinical staff make the TCM interactive contact within 2 business days and complete medication reconciliation, and agents route any clinical question to them by rule.
For CCM billing, AI agents conduct monthly check-in calls, route reported symptom changes to care managers by rule, track medication adherence, and prepare notes so clinical staff can document the 20 minutes of clinical staff time per calendar month that CPT 99490 requires. This helps practices capture CCM revenue at scale without a matching increase in staff.
AI identifies patients with open quality measures (HEDIS, STARS, HCC gaps), prioritizes by impact and likelihood of closure, and conducts automated outreach to schedule the necessary visits, screenings, or assessments.
Start with your highest-volume, most standardized workflow โ typically TCM or appointment reminders. Demonstrate results on that workflow against a measured baseline. Then expand to chronic care management, care gap closure, and specialty-specific protocols. The key principle: AI handles the volume so your care managers can handle the complexity.
Medicare's care management codes define who must do which part of the work, so they set the boundary for automation. Agents can carry the outreach and scheduling around each service; the billable clinical parts stay with licensed staff.
For transitional care management, the physician or clinical staff under their direction must contact the patient or caregiver within 2 business days of discharge, the contact must be made by clinical staff who can address the patient's status, and medication reconciliation must happen on or before the face-to-face visit, which is due within 7 or 14 calendar days depending on complexity. For chronic care management, eligible patients have two or more chronic conditions expected to last at least 12 months, and CPT 99490 covers the first 20 minutes of clinical staff time in a calendar month, provided under the billing practitioner's general supervision.
CCM minutes count time spent by the billing practitioner or clinical staff, so time an agent spends on a call does not count toward them. What agents can do is make the staff minutes count: reach the patient, collect the check-in answers, flag changes and prepare notes, so the care manager's time goes to the clinical conversation. Care plans for TCM, CCM and wellness visits can sequence the agent steps and the staff steps in one workflow.
Measure the work that moved off your care managers and the outcomes that depend on it, against a baseline taken before you start.
For TCM, track discharges detected, patients reached, interactive contacts completed within 2 business days and face-to-face visits held inside the window. For CCM, track enrolled patients with a completed monthly check-in, escalations to care managers and how quickly they were resolved, and months that met the time requirement. For care gaps, track gaps worked, visits booked and gaps closed. Then look at care manager time: how many hours shifted from dialing and paperwork to complex patients.
Keep a short list of what agents must always hand to a person: symptom changes, medication problems, crisis signals, caregiver confusion and any request to speak with someone. Our guide to how AI agents work with your care team shows how those handoffs are set by rule.
Agents handle the repeatable, rules-based work: initial and follow-up outreach calls, appointment scheduling, medication and visit reminders, barrier screening, transportation coordination, referral tracking and preparing documentation of standard interactions. They work within the rules and hours your team sets and route symptom changes, medication problems and anything that needs clinical judgment to a care manager.
No. CMS counts CCM time spent by the billing practitioner or by clinical staff under their general supervision, and CPT 99490 covers the first 20 minutes of clinical staff time in a calendar month. An agent's call time is not clinical staff time. Agents can collect check-in answers and prepare notes so that staff minutes go to the clinical conversation, and the billing practitioner should review the documented time before billing.
Start with the highest-volume, most standardized workflow, usually transitional care management after discharge or appointment reminders. Define what the agent does and what it hands to a person, take a baseline, run it with one team and compare results after a few months. Then expand to chronic care management, care gap closure and specialty protocols. ADT-triggered TCM outreach is a common first workflow for ACOs.
The goal is capacity, not a smaller team. Automation moves dialing, scheduling and routine documentation off care managers so they can spend more time on complex patients, care plans and conversations that need clinical judgment. Panels can grow without each care manager drowning in administrative work, but every clinical decision, escalation and billable clinical service still needs a licensed person.
CMS Medicare Learning Network: Transitional Care Management Services (MLN908628). https://www.cms.gov/files/document/mln908628-transitional-care-management-services.pdf
CMS Medicare Learning Network: Chronic Care Management Services (MLN909188). https://www.cms.gov/files/document/chroniccaremanagement.pdf
Jay Chowdappa, MD is co-founder and CEO of Zynix AI. He is a physician, and ACOs led by him generated $300M+ in shared savings before he started Zynix AI to help care teams follow through on what their data already shows.