Health Plans Deploy Artificial Intelligence to Preserve Revenue When Medicaid Renewals Shift to Twice Yearly

Kern Family Health Care in California replaced what would have been 40 full-time workers with an AI agent named Angelica, spending $370,000 on software that has placed more than 800,000 calls to 387,000 Medi-Cal members since late 2025 . The timing matters: mandatory Medicaid work requirement documentation takes effect nationally in 2027, and the One Big Beautiful Bill Act signed by President Trump in 2025 now requires most Medicaid patients to renew enrollment twice yearly instead of annually . For health plans paid through managed care contracts, keeping people enrolled is revenue preservation. For healthcare marketers, this is a preview of how patient engagement will shift when compliance burdens double and labor costs become unsustainable.

The financial calculation is stark. Kern Family Health Care, the largest Medi-Cal provider in Kern County where 52% of residents rely on the safety net program, estimated the AI program saved $2.4 million in staffing costs . The software, developed by San Francisco startup Careforce, conducts outreach in patients' preferred languages, schedules appointments with human staffers who verify paperwork, and reschedules when conflicts arise. Vanessa Barahona, a 41-year-old office cleaner, completed her renewal with "Angelica" and told KFF Health News through a translator: "It felt like I was talking to a real person" . She wasn't. And that disconnect — between patient perception and operational reality — is where healthcare marketing enters legally and ethically complex territory.

Emily Duran, CEO of Kern Health Services, said the plan worked with Kern County's Department of Human Services to obtain data that allows Kern Family to know when a member's eligibility will expire . The county stations workers inside Kern Family's Bakersfield facility to answer enrollment questions. This operational model — AI for outreach volume, humans for verification — mirrors what Baltimore deployed in 2026 for mental health crisis response, where mobile teams of clinicians and peer counselors now handle calls previously routed to police . Both models reflect the same strategic shift: technology handles scale and initial contact; humans handle complexity and trust.

The Economics of Automated Enrollment Under New Federal Rules

The managed care payment model creates direct financial incentive for health plans to keep members enrolled. When patients lose coverage, plans lose revenue. Kern Family's $370,000 investment in Careforce software delivered 800,000 outreach calls across 387,000 members — roughly 2.1 calls per member since late 2025 . With twice-yearly renewals beginning in 2027, that call volume will double. The alternative — hiring 40 full-time employees at an estimated $2.4 million annually — makes the business case clear .

Mark Duggan, a Stanford University economics professor who has studied Medicaid for 30 years, said one long-standing fear is that insurers could use such software to cherry-pick patients for coverage . The concern is not hypothetical. Health plans earn capitated payments per member per month, but not all members generate the same margin. An AI system optimized for enrollment efficiency could, in theory, prioritize healthier, lower-cost patients for outreach while allowing higher-utilizers to lapse through administrative friction. Kern Family officials say no workers have lost jobs, and the plan is not unionized . But the labor calculus matters for marketers: when patient engagement infrastructure shifts from humans to software, trust becomes a marketing asset, not an HR line item.

Anthony Cava, spokesperson for California's Department of Health Care Services, said Medi-Cal health plans have flexibility in how they handle member renewals, including with AI tools, but plans remain responsible for ensuring technology complies with state and federal regulations, including patient privacy and data security . The department began allowing managed care plans to contact members about renewals last year, foreseeing the huge volume of reenrollments required statewide. State rules still prohibit Medi-Cal health plans from soliciting new enrollees — only county health officials determine eligibility . That regulatory boundary is critical for marketers: outreach for retention is permissible; outreach for acquisition is not.

When Patients Don't Know They're Talking to Software

Vanessa Barahona's experience reveals both the promise and the risk. She completed her renewal, avoided a coverage gap, and appreciated that "Angelica" spoke Spanish. She also believed she was talking to a real person . That perception gap is not unique to Medicaid enrollment. Behavioral Health System Baltimore oversees mobile crisis teams that have diverted mental health calls from police to clinicians for years, but a 38-year-old woman who called 911 in March 2026 while in crisis received a human clinician response — and told the clinician that a previous attempt to get help ended with her arrest . The distinction between AI-assisted triage and human clinical judgment matters when the stakes are health outcomes, not just administrative convenience.

Careforce CEO Huzaifa Sial said Kern Family is one of a few health insurers using his company's software to boost Medi-Cal enrollment, and the company is also working with the Central California Alliance for Health . Sial framed the value proposition as execution: "Most people don't know what they need, and if they do, they have a hard time getting there. That's the hidden execution problem that nobody sees" . That framing treats patient engagement as a logistics problem. But healthcare marketing leaders understand that trust is not logistical. When a patient believes they are speaking to a person and discovers they were not, the trust calculation changes — and with it, the likelihood of future engagement.

The rise of AI in healthcare has prompted concerns about improper denial of treatments, workplace surveillance, algorithmic biases, transparency, data privacy, and safety risks . Unions have raised concerns about prioritizing savings over safety. Polling shows widespread fear over AI-driven job losses and growing income inequality . For health plans, these concerns are reputational risks. For healthcare marketers, they are positioning challenges. How do you message efficiency gains from automation without eroding the human connection that drives patient loyalty?

Regulatory Oversight Lags Behind Deployment

California health plan regulators say they are tracking AI use closely, and the state attorney general's office has sent reminders to healthcare entities about their obligation to follow consumer protection rules . But the oversight framework remains reactive. The FDA announced in July 2026 that it was revoking outdated authorizations for petroleum-based color additives in food, part of a broader "Make America Healthy Again" regulatory reform agenda . That same principle — updating regulations to reflect current practices — has not yet been applied systematically to AI in patient engagement. The Office of the National Coordinator for Health IT released 2026 approved Standards Version Advancement Process (SVAP) standards focused on interoperability and data exchange , but those standards do not address conversational AI transparency or patient notification requirements.

The regulatory gap creates risk for health plans and opportunity for marketers who position their organizations as transparent early adopters. Kern Family worked closely with Kern County to station government workers inside its Bakersfield facility . That visible partnership signals legitimacy. Vanessa Frando, chief deputy director of Kern County Human Services, called Kern Family's leadership "very innovative and forward-thinking" . That institutional endorsement matters when patients are skeptical of automation.

The 1ness Take

Healthcare marketers face a strategic choice: deploy AI for patient engagement efficiency and risk trust erosion, or invest in hybrid models that preserve human touchpoints and differentiate on transparency. The Kern Family case proves the economics work. The Barahona experience proves patients cannot reliably distinguish AI from humans. The regulatory environment proves oversight is lagging. That combination creates a narrow window for competitive advantage.

Our recommendation: treat AI disclosure as a marketing asset, not a compliance burden. Health plans and provider organizations that proactively notify patients when they are interacting with AI, explain what the AI can and cannot do, and offer clear pathways to human escalation will differentiate themselves as patient-centric at a time when competitors are optimizing for cost. Kern Family saved $2.4 million in staffing costs, but that savings will evaporate if patient trust declines and disenrollment increases once the twice-yearly renewal requirement begins in 2027.

Build marketing messaging around the hybrid model: "Our AI handles scheduling and reminders in your language, 24/7. Our team handles everything that requires a decision." That positions automation as augmentation, not replacement. It also creates a compliance narrative for regulators who will eventually require transparency. The health systems that establish trust-forward AI engagement now will own the patient loyalty advantage when federal work requirements double the renewal burden in 2027.

For patient acquisition, the constraints are tighter. California prohibits Medi-Cal plans from soliciting new enrollees — only counties determine eligibility . But retention marketing is permissible, and retention is where the AI economics are most compelling. An engagement strategy that uses AI for outreach scale and humans for trust-building can reduce disenrollment while preserving margin. That requires content marketing that educates patients on how to navigate renewals, multi-language support that reflects community demographics, and a visible partnership with county agencies that signals legitimacy. Kern Family's model — county workers stationed inside the health plan facility — is operationally smart and optically strategic.

Finally, track your AI engagement metrics separately from human engagement metrics. Measure not just call completion rates but also escalation rates, patient satisfaction scores by channel, and disenrollment rates by engagement type. If AI-engaged patients disenroll at higher rates than human-engaged patients, the cost savings are illusory. If they disenroll at similar or lower rates, you have a data-driven case for expanded automation. Either way, you have a defensible answer when regulators, journalists, or community advocates ask whether your AI is helping patients or just cutting costs.

The Takeaway

Healthcare marketers navigating the shift to AI-assisted patient engagement under new federal Medicaid rules should take three specific actions:

• Audit your patient communication technology for transparency gaps now. If your health plan or provider organization uses conversational AI for enrollment, appointment scheduling, or care reminders, determine whether patients are notified they are interacting with software. If not, build the business case for proactive disclosure as a trust-building differentiator before regulatory requirements force reactive compliance.

• Separate retention marketing from acquisition marketing in Medicaid contexts. State rules vary, but California's prohibition on health plan solicitation of new Medi-Cal enrollees is instructive . Build content strategies, community partnerships, and multi-language support around helping existing members navigate renewals successfully. That investment pays off twice: in reduced disenrollment and in positioning your organization as patient-focused when competitors are optimizing only for cost.

• Model the economics of twice-yearly renewals before 2027. The One Big Beautiful Bill Act doubles the compliance burden for most Medicaid patients beginning in 2027 . If your engagement infrastructure cannot handle that volume without proportional cost increases, you will face the same build-versus-buy decision Kern Family faced. Run the numbers on AI augmentation now, while you still have time to pilot, measure, and refine before the regulatory deadline.

---