By Jay Chowdappa, MD, co-founder and CEO, Zynix AI · Updated November 4, 2026
Key takeaways
ACOs can scale operations without hiring in step with panel growth by automating the high-volume, rules-based work that grows with every new attributed patient: unified data, risk and gap prioritization, outreach and scheduling by AI agents, and care plans that sequence each program. Staff then spend their time on complex patients and clinical decisions instead of lists and phone queues.
- Manual TCM outreach, AWV scheduling and gap lists are the workflows that break first as an attributed panel grows.
- A single reconciled patient record removes the need for staff to work across separate systems.
- Prioritized worklists focus effort on the highest-impact gaps instead of the easiest ones.
- Agents carry outreach, scheduling and reminders, while licensed clinicians make TCM contacts and clinical decisions.
- Measure completed work per staff member, not call volume, to know whether the model is scaling.
The ACO Scaling Challenge
Many ACOs hit a scaling wall as their attributed population grows. The workflows that worked for a smaller panel — manual outreach, spreadsheet-based tracking, ad hoc care coordination — collapse under the weight of larger panels. Hiring proportionally more staff erodes shared savings margins. The organizations that break through this wall do so by automating their operational backbone.
Where ACOs Lose Efficiency at Scale
TCM follow-up: As monthly discharges climb into the hundreds, a manual TCM workflow needs a growing team just for initial outreach. When outreach can't keep pace with the two-business-day contact window, contact rates fall and TCM revenue goes uncaptured.
AWV scheduling: Scheduling thousands of AWVs a year through a call center is expensive per visit booked, and many eligible patients still go without one.
Gap closure: With thousands of open HCC and quality gaps, manual prioritization and outreach becomes impossible. Staff default to working easy-to-close gaps rather than highest-impact gaps.
The AI-First ACO Operating Model
Layer 1: Unified Data Foundation
Connect every data source — EHRs, claims, ADT feeds, labs, pharmacy — into a single reconciled patient record. This eliminates the data fragmentation that forces staff to toggle between multiple systems.
Layer 2: Intelligent Prioritization
AI risk models and gap detection engines continuously analyze the full panel, surfacing exactly who needs what intervention and when. Staff work from dynamically prioritized worklists rather than static reports.
Layer 3: Automated Execution
AI agents handle the volume work: TCM outreach calls, AWV scheduling, appointment reminders, care gap outreach, and gathering medication lists for clinician reconciliation. Licensed clinicians make the TCM interactive contact, and human staff focus on complex cases requiring clinical judgment.
Layer 4: Orchestrated Care Plans
Multi-step care programs (TCM 30-day, chronic care, gap closure sprints) are orchestrated by AI — automatically sequencing outreach, tracking milestones, escalating exceptions, and documenting outcomes.
The Math That Works
An ACO using AI-supported operations aims for more discharged patients reached inside the TCM window, more completed AWVs, more gaps closed, and less staff time on routine tasks. The result: better quality scores, more shared savings, and a scalable foundation for growth.
What to automate first
Start where the work is high-volume, rules-based and time-bound, because those steps break first as the panel grows. For many ACOs that means post-discharge outreach, wellness visit scheduling and reminders.
Post-discharge outreach has a hard clock: Medicare's TCM rules call for an interactive contact within 2 business days of discharge and a face-to-face visit within 7 or 14 days, depending on complexity. Wellness visits and gap closure are less urgent but much larger in volume. In both, agents can carry the calls, texts, booking and reminders while nurses make the clinical contacts. ADT-triggered TCM outreach is a common first workflow because its deadlines make results easy to measure.
How to tell the model is working
Measure completed work per staff member, not activity. The question is whether the same team finishes more of the work that drives quality and savings.
- Share of discharges with an outreach attempt and a clinical contact inside the TCM window.
- Annual wellness visits completed against the eligible list, by practice.
- Open quality and HCC gaps closed before year-end, and whether they were worked by priority rather than ease.
- Staff hours spent on routine calls and list building compared with time on complex patients.
- Escalations from agents to staff, and how quickly each one was resolved.
Frequently asked questions
How can an ACO grow its attributed lives without hiring proportionally more staff?
Automate the work that grows in step with the panel: finding discharges, building outreach lists, calling and texting patients, booking visits and sending reminders. Put all data into one patient record so staff stop reconciling spreadsheets, rank patients by risk and open gaps, and let agents carry routine outreach within rules your team sets. Staff time then goes to complex patients and clinical work.
Which ACO workflows should be automated first?
Start with workflows that are high-volume, rules-based and on a deadline. Post-discharge TCM outreach is a common first choice because the contact window is short and results are easy to track. Annual wellness visit scheduling and year-end gap outreach follow, since they involve many patients and mostly logistics. Keep clinical steps, such as the TCM interactive contact and medication reconciliation, with licensed staff.
What results should an ACO expect from AI-supported operations?
Expect results to be measured on your own data rather than promised up front. Set a baseline before launch for TCM contact timing, AWV completion, gaps closed and staff hours on routine work, then compare after a full cycle. Zynix customer stories describe how ACOs, health plans and physician groups have set up these programs.
What stays with human staff when agents handle ACO outreach?
Clinical judgment and relationships. Nurses and physicians make the TCM interactive contact, reconcile medications, assess symptoms and decide on treatment, and care managers handle complex patients and difficult conversations. Agents route symptom questions, barriers and exceptions to the right person by rule and never diagnose, triage clinically or recommend treatment.
The 30/60/90-day ACO scale-up framework
Scaling ACO operations with AI is not an overnight event. Here is what a successful 90-day program ramp looks like, with concrete milestones at each gate.
| Milestone | Days 1–30 | Days 31–60 | Days 61–90 |
|---|---|---|---|
| Data infrastructure | Live EHR and ADT connection; initial patient panel loaded and validated | Claims data integrated; risk model running on full panel | Quality measure gaps calculated; benchmark data imported |
| Workflow automation | TCM outreach agent running; first discharges contacted within window | AWV scheduling live; second workflow fully staffed and monitored | Gap outreach (HCC, HEDIS) live; all three major workflows in production |
| Staff adoption | Care managers using embedded worklist daily; outliers reviewed weekly | Physicians reviewing agent-generated care plans; escalation queue under 10% of contacts | Program managers reporting on workflow metrics; QI loop closed |
| Financial metrics | TCM billing capture baseline established | TCM capture rate trending up; AWV completion rate tracked | Month 3 quality measure gap closure visible in dashboard; first shared savings estimate updated |
Related reading
- How ACOs Actually Make Money: Shared Savings Mechanics and Where AI Moves the Needle
- How To Automate Care Management With AI Agents
- Annual Wellness Visits: How ACOs Can Build the First Layer of Scalable Care Capacity
Source notes
CMS (Medicare Learning Network): Transitional Care Management Services, MLN908628 (August 2025). https://www.cms.gov/files/document/mln908628-transitional-care-management-services.pdf
About the author
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.