A US-based managed care organization serving Medicaid and Medicare members needed to fix a growing operational bottleneck: assigning and re-assigning members to Care Managers (CMs) at scale. With 2M+ members and 4,500+ field CMs across states, spreadsheet-driven workflows and basic system rules couldn’t reliably factor in travel time, member acuity, or caseload balance. The result was delayed outreach for high-priority members, inconsistent service levels, and higher operating costs driven by inefficient routing and uneven workloads—some CMs overloaded while others were underused. The client partnered with Persistent to implement an AI-driven optimization engine to automate assignments and improve field efficiency.
Where Wheels Wobbled: Assignment Pains & Travel Drains
The client discovered repeated delays in first outreach for certain high-risk, long-term care cohorts, despite increasing CM headcount. Leadership recognized that adding more staff alone would not solve the problem; the issue lay in how members were being assigned and how CMs’ time was being used.
The following structural challenges affected assignments, field workflows, caseload balance, scalability, and service quality.
- Manual, time-consuming workflows: Member assignment and re-assignment relied on manual spreadsheets and informal rules, with regional staff matching members to CMs and entering updates into the platform. This caused outreach delays, higher assignment and poor scalability as membership and program complexity increased.
- Inefficient travel planning for field-based care managers: Assignments did not reflect real drive times or routes, forcing CMs to travel long distances with inefficient, manual scheduling. This caused backtracking, idle time and reduced availability for member contact, documentation and care coordination, directly delaying time-to-visit, which could have been particularly harmful for high-acuity and post-discharge members.
- Resource optimization gaps in long-term care: With complex long-term care programs, the members needed recurring visits with strict timeframes; caseloads had to reflect acuity, service intensity and visit frequency, not just counts. Lacking data-driven caseload balancing, the client faced growing operational challenges.
- Caseload imbalances: Some CMs managed disproportionately more high-acuity members than others, causing underutilized CM capacity in certain regions and raising the risk of missed visit windows and unmet service-level expectations.
From Spreadsheets to Smart Routes: Assignments That Find Their Way
Persistent designed and implemented a full-fledged AI-driven optimization solution for member assignment and re-assignment, with a focus on route optimization, caseload balancing and operational enablement.
Our data experts helped the client build a strong data foundation by combining member data, such as demographics, geo-coordinates, enrolled programs and risk scores. We linked this with CM data, such as work location, credentials, caseload and acuity mix, and overlaid this with operational data, such as time to first contact, reassignment reasons, and travel history in a central SQL data warehouse with quality checks and standard geo-coding.
The data layer set the stage for AI-based decisioning and optimization around routing and constraint-based assignment. The AI model uses a mix of heuristic and mathematical optimization to minimize total travel time and distance, while a constraint-based logic ensures parity among CMs regarding caseload, visit frequencies, specialized-skill matching and thresholds for geographic coverage and travel radius.
To ensure the optimization engine could be trusted in day-to-day operations, we validated integration touchpoints using SOAPUI, automated end-to-end regression testing of critical workflows with Leapwork, and generated synthetic, PHI-safe test data with GenRocket. This strengthened release confidence while protecting sensitive member information. Once validated, assignment recommendations surfaced directly in the care management platform as actionable worklists that supervisors could review and see the key drivers behind each recommendation (for example, travel time, acuity and caseload constraints), approve or override with reason codes, and maintain a complete audit trail for compliance.
With AI-led, automated assignment runs, CMs receive pre-optimized daily routes and prioritized visit lists tailored to their caseload and member needs, and reassignments due to CM leave or coverage changes are automatically suggested instead of manually recalculated. This shift has reduced administrative time spent on assignment decisions and freed up leadership to focus more on care quality and program management.
Vitals Up, Miles Down: Healthier Routes, Happier Members
Within the initial deployment (covering a subset of regions and LTC programs over approximately 16 weeks from design to pilot), the client observed:
- 2× faster care delivery enabled by reduced travel: Average daily travel per CM in pilot regions decreased from 21 km to 9.5 km. This effectively doubled the available time for member contact and care coordination activities, translating into faster scheduling of initial and follow-up visits, and improved adherence to regulatory timelines in LTC programs.
- Reduced cost through more efficient allocation: Lower travel time and better caseload distribution reduced overtime and ad hoc staffing needs. The client reduced operational cost per completed visit (directionally informed by fewer travel hours and more visits completed per CM per day). Coordinators and supervisors spent significantly less time on manual assignment tasks, enabling them to support more members and programs without adding headcount.
- Enhanced operational efficiency and resource optimization: Caseload variance across CMs was narrowed into a defined equitable band, improving fairness and reducing burnout risk. LTC programs benefited from more predictable coverage, with improved on-time visit completion rates and fewer last-minute reassignments. Improved visibility into assignments and routes allowed leadership to plan regional staffing strategies based on real data rather than anecdotal feedback.
- Improved experience indicators: Fewer complaints related to delayed outreach or missed visits and positive qualitative feedback from CMs regarding more manageable and predictable days.
By modernizing member-to-CM assignment, route planning and workload balancing with an AI optimization engine, the client turned a long-standing operational pain point into a scalable capability that supports timely, equitable and efficient care, particularly for its most vulnerable LTC members.
