This paper is part of CHCF’s Bold Ideas for Medi-Cal initiative, which sought innovative ideas for strengthening California’s Medicaid program. The views expressed are those of the authors and do not necessarily reflect the views of the California Health Care Foundation.

Executive Summary

California has a once-in-a-generation opportunity to redesign Medi-Cal as a more preventive public service with AI. Despite major investments through CalAIM (California Advancing and Innovating Medi-Cal), the current system remains relatively reactive rather than proactive. With the use of AI, risk can be mitigated upstream, before a member presents in the emergency department, loses housing, or cycles through disconnected programs.

We propose that California build on existing infrastructure to create an AI “app store” for Medi-Cal — one where members choose to work with AI Advocates developed by state agencies, health plans, and certified technology vendors to help them manage their health, benefit, and social service needs. Rather than waiting for a crisis, these AI Advocates would proactively respond to rising risk levels and initiate outreach, navigation, and enrollment. The approach would include privacy and security safeguards, human oversight, consent management, optional delegated access, and equitable support across voice, text, and secure mobile and web apps.

This strategy follows a public-private model: The state builds and governs the enabling infrastructure, and qualified vendors build and operate the AI Advocates that serve members. With informed member consent managed by the state, AI Advocates would access Medi-Cal member data through Medi-Cal Connect, the Population Health Management Service operated by the California Department of Health Care Services (DHCS). Critically, this approach would not require consolidating data into a single system; instead, data would become accessible to authorized AI Advocates through the Model Context Protocol, an open-source AI standard. With the foundation of existing state infrastructure and transparent governance, state-certified AI Advocate applications would compete on user experience and outcomes while the state managed consent and ensured strong human oversight.

This approach would enable data-driven, member-centered coordination that reduces preventable emergency department visits and hospitalizations; increases access to services that address housing, justice, and behavioral health needs; and reduces administrative burden and handoffs across counties, providers, and community organizations.

A $5 million to $10 million demonstration (eligible for federal matching funds) would validate the approach and improve outcomes at scale while reducing avoidable high-cost crises. Federal policy changes and rising health care costs heighten the urgency for a predictive, preventive, and person-centered Medi-Cal system enabled by AI. Although built on California’s existing data infrastructure, this open-standard approach offers a replicable model for AI-native Medicaid platforms nationwide.

The Problem

Despite innovative investments in new care modalities and access to social care integrated into health systems, the Medi-Cal program faces numerous challenges in meeting member needs at the right time, in the right place, with the right level of service. These barriers include the following issues:

• Fragmentation, late intervention, and low use of preventive care. Health care, behavioral health, benefits, housing, education, justice, and other data sources are siloed within state and county systems. While segmentation supports privacy and alignment with applicable laws, it also limits the use of existing data to coordinate person-centered care and reduce costs. Indicators of rising risk for Medi-Cal members may appear in existing data systems months or years before crises; however, the delivery system often intervenes only after conditions worsen.^[i]
  
  Through CalAIM (California Advancing and Innovating Medi-Cal), the Behavioral Health Services Act, the Birthing Care Pathway, and other statewide initiatives, California has expanded both the services available to Medi-Cal members and the data infrastructure that supports them.^[ii] Medi-Cal Connect launched in 2024 as the statewide Population Health Management Service for Medi-Cal, with the California Department of Health Care Services (DHCS), managed care plans (MCPs), and care delivery partners gaining access in phases.^[iii] Members and their caregivers, however, have no access. The state now possesses more tools and actionable information than ever. Yet use of preventive, lower-cost services remains uneven, while crisis-driven interventions continue to consume a disproportionate share of spending.^[iv] For example, CalAIM Enhanced Care Management (ECM) use currently falls short of the estimated 3% to 5% of eligible members.^[v]
• Administrative burden and workforce constraints. Medi-Cal eligibility determination, redeterminations, documentation, member outreach, and care coordination require sustained hands-on support across fragmented systems. These administrative functions are essential, but they are resource-intensive and increasingly strained, particularly across primary care, behavioral health, and direct care workers.^[vi] Recent federal changes under H.R. 1 will intensify these pressures.^[vii] For members, more stringent qualification requirements will most likely mean dropped enrollment. Fewer qualified members may result in decreased funding to health plans and providers, reducing the very workforce capacity that is needed to manage heightened administrative demands.
• Equity and member experience failures. Language barriers, complex rules, and opaque navigation disproportionately burden communities who are facing poverty, disability, racism, immigration-related issues, and other inequities. In California, roughly 44.4% of residents speak a language other than English at home.^[viii] Workforce data show that even with over 50 languages represented among licensed health professionals, key languages such as Spanish are underrepresented relative to population language prevalence, highlighting a gap in linguistic and cultural workforce capacity.^[ix] The result is delayed care, unmet social needs, and preventable crises. Federal policy changes that increase redetermination frequency and introduce work requirements will further heighten churn and procedural terminations. These shifts risk deepening mistrust and disengagement among the members who stand to benefit most from California’s expanded services. Without structural simplification and proactive support, administrative complexity will continue to function as a barrier to equity rather than a pathway to coverage.

If the status quo persists, Medi-Cal faces rising long-term costs, widening inequities, and erosion of the gains achieved through CalAIM and related reforms. Without a transformative approach to early intervention and proactive member engagement, the program will struggle to realize the full promise of its recent investments.

The Bold Idea

Core Concept

California’s health care system has historically responded after harm has occurred. Medi-Cal members often come into view when they seek emergency care, miss renewal paperwork, or experience destabilizing life events. Data that could signal impending crises exist across programs, but the data are siloed and are rarely used in time to prevent deterioration. California can shift to a more coordinated, preventive approach by establishing AI Advocates for all Medi-Cal members, built to integrate directly with Medi-Cal Connect.

AI Advocates would create a unified, member-facing experience by leveraging existing state assets, including Medi-Cal Connect and the consent management platform, without the need for new databases or aggregation. Consent for an AI Advocate to access a member’s data in Medi-Cal Connect would be obtained during Medi-Cal plan enrollment or redetermination. Consent could be revoked at any time. In Medi-Cal Connect, AI Advocates would access medical, behavioral health, and pharmacy claims; provider information; ECM and Community Supports enrollment status; and risk stratification. With these data, AI Advocates could detect utilization and eligibility changes, address rising clinical or behavioral health risk, and prompt tailored outreach while facilitating enrollment in appropriate services. To broaden access to 24/7 support, AI Advocates would interact with members in their preferred language via text, voice, or a mobile or web portal. Medi-Cal would function as a member-centered system that anticipates needs, alerts members, and connects them to support before conditions worsen. This innovation would transform how Medi-Cal members interact with the health care system.

Future expansion of the initiative to enable AI Advocates to access additional data sources, including real-time clinical data, social services, housing, education, employment, and public utilities, would enhance coordination capabilities across siloed health and social sectors.

Transparent governance and human oversight would be prioritized at every step. While AI Advocates would support members in identifying appropriate benefits and navigating to enrollment, they would not supplant the existing person-to-person services that MCPs, counties, and providers offer to members, and they would not replace clinical roles. AI Advocates would refer members to ECM providers and other case managers as appropriate by using the provider information available in Medi-Cal Connect.

Member and System Impact

Medi-Cal members who experience destabilizing life events, such as childbirth, job loss, housing instability, and new diagnoses, face heightened risk of coverage loss and care disruption while administrative requirements simultaneously intensify. AI Advocates could detect relevant signals in available data, alert members to services that they qualify for, and guide members through renewals or enrollment in services. This is particularly important for members facing language barriers, disability, or homelessness, for whom fragmented systems exacerbate inequities. To address the churn anticipated by H.R. 1, DHCS may consider expanding the service to previous members who need support in reenrolling.

With AI Advocates checking eligibility for members, providers and community-based organizations would spend less time verifying eligibility and tracking documentation across disconnected systems. The limited workforce capacity could then focus on care delivery rather than administrative tasks. Counties, MCPs, and state agencies would see improved retention, fewer avoidable coverage gaps, and more timely identification of members who need support. By translating early signals in state data systems into timely outreach and enrollment support, AI Advocates would help shift Medi-Cal from a crisis-driven system to one that intervenes earlier and maintains continuous coverage.

How It Would Work in Practice

Because California has already built much of the initial infrastructure, success of the AI Advocates would depend on aligning the existing statutory authority, financing pathways, and data systems to deliver a coordinated, member-facing capability.

Program Architecture

Figure 1. AI Advocate Architecture, with an Example Member

Source: BluePath Health, 2026.

The AI Advocates would be initially designed to access data in Medi-Cal Connect. This model requires the state to establish a consent management strategy for AI Advocates. State-certified technology vendors would then build member-facing AI Advocates to securely access Medi-Cal Connect, with the Model Context Protocol adopted to enable standard integration.

Consent management platform. DHCS is actively developing a consent management platform to store and to manage standardized consent records. Building on the Authorization to Share Confidential Member Information (ASCMI) initiative, the platform supports identity matching and secure storage of member authorizations that can be accessed by approved intermediaries.^[x] To enable AI Advocates, the consent management platform would also manage member consent for AI Advocates, collected at the point of Medi-Cal plan enrollment or redetermination. Consent would be explicit, revocable, and auditable. Members could also designate a delegate, such as a family member or care team representative, to receive communications and to interact with an AI Advocate on their behalf.

Model Context Protocol for integration. This initiative proposes adopting the Model Context Protocol, an open-source standard for connecting AI systems to external data sources, as the statewide integration standard.^[xi] Under this model, the state would build a standardized server endpoint in Medi-Cal Connect to serve as a secure, structured interface through which the system could be queried by authorized AI Advocates while data remain where they already live.

Through this federated architecture, the state retains control of its own systems and data. The state would also avoid the cost and risk of building a centralized data warehouse. Because the Model Context Protocol is an open-source standard, it would prevent vendor lock-in and enable a competitive marketplace. Standardized server endpoints would enable future expansion of the AI Advocates initiative to access other existing databases.

AI Advocates (vendor built). The state would establish a regulatory framework and certification system for AI Advocate technology vendors. Certified vendors would build AI advocates that use consent-mediated access to Medi-Cal Connect to support risk tracking, eligibility monitoring, and member outreach. After members have opted in, AI Advocates would analyze their available claims, eligibility, and administrative data to identify potential eligibility for additional services based on the Medi-Cal Connect Risk Stratification, Segmentation, and Tiering (RSST) algorithm.^[xii]

The AI Advocates would monitor for indicators such as new diagnoses, ED visits, upcoming renewals, pregnancy, or lapses in benefits. When defined thresholds have been met, it would initiate proactive outreach in the member’s preferred language and through their preferred channel, explain available options in plain terms, and direct the member to the appropriate enrollment pathway, with streamlined escalation to human support.

Vendors would compete in an AI Advocate “app store” based on the quality of the user experience, including the effectiveness of outreach, multilingual capability, clarity of guidance, and outcomes for members. The state would set the standards for privacy, safety, consent, data access, and clinical appropriateness, ensuring that storage of any information within AI Advocates is compliant with the Health Insurance Portability and Accountability Act, the Confidentiality of Medical Information Act, and other applicable state and federal laws.^[xiii]Data sovereignty requirements will be central to certification of AI Advocates, ensuring that members retain full ownership of their data.

Member Experience

Participation would be voluntary. Members would learn about AI Advocates during Medi-Cal enrollment and redetermination, with education supported by MCPs, county agencies, medical providers, 2-1-1 navigators, and other trusted community partners. Educational materials would clearly explain the purpose of the tool; the scope of data access; privacy and security considerations; human oversight; digital literacy and fraud prevention practices; and the member’s right to decline, withdraw consent at any time, or select a delegate. Members could interact with AI Advocates through voice calls and text over standard cellular networks to ensure core accessibility without requiring app downloads, smartphones, or broadband. A secure, web- and mobile-accessible portal would also be available for members who prefer digital self-service. Outreach would occur in the member’s preferred language across all channels. The interface would provide clear explanations of benefits, renewal timelines, and next steps, reducing reliance on capacity-strained and fragmented agency contacts.

Member-centered design is a core principle of the AI Advocates. Codesigning with member input and improvements driven by member feedback are essential components of the development process that would strengthen trust and utilization.

Governance

DHCS would oversee governance and stewardship of the AI Advocates, building on its leadership of Medi-Cal Connect and ASCMI. DHCS would establish standards for privacy, security, certification, and responsible AI use to ensure that AI Advocates operate safely and equitably. By developing AI Advocates as an enhancement to Medi-Cal Connect and by integrating functionality into the consent management platform, DHCS could draw down 50% federal matching funds through Medicaid Administrative Claiming for the eligible administrative system enhancements.^[xiv]

In the 2021–22 California state budget, the Population Health Management Service, later named Medi-Cal Connect, was introduced as a tool to “facilitate the connection between important health data and critical human service data for a given beneficiary,” a critical building block for this initiative.^[xv] Under CalAIM, legislation was also signed in 2021 to compel “the sharing of health, social services, housing, and criminal justice information, records, and other data.”^[xvi] These CalAIM authorities could support the legal foundation and initial pathway necessary to enable AI Advocates. For future expansions of AI Advocates to access other data systems, additional legislation or executive orders would be necessary.

Certified technology vendors would build and operate the member-facing AI Advocates within a DHCS-administered certification framework, competing on accessibility and the quality of the member experience while adhering to state-defined safety and privacy requirements. If the service expanded to county-governed eligibility and public benefits systems, counties would play a governance role. Coordination with the California Statewide Automated Welfare System (CalSAWS) would be necessary to enable secure data access and operational integration, likely requiring legislation or an executive order. Community-based organizations and 2-1-1 networks would support outreach and escalation of member needs that were not resolved digitally.

In future expansions of the initiative, access to real-time clinical data would be enabled through the Data Exchange Framework Qualified Health Information Organizations (QHIOs) that exchange Medi-Cal member data.^[xvii] DHCS’s existing contracting authority with MCPs provides an appropriate mechanism to ensure consistent clinical data sharing across the state. Through their contracts, DHCS requires MCPs to execute and maintain participation in the Data Sharing Agreement, which many MCPs fulfill through QHIOs. DHCS could extend this contractual requirement for MCPs to exchange data through a participating QHIO. To support standardized, secure access, DHCS could require MCPs to ensure that their QHIOs expose approved interoperability endpoints that can support AI Advocates by using the Model Context Protocol. AI Advocates could then retrieve relevant clinical context in a governed and auditable manner without requiring integration into each MCP’s internal systems.

Implementation Pathway

AI Advocates would first be built as a component of Medi-Cal Connect, leveraging the existing claims data, eligibility information, and risk stratification capabilities. DHCS would incorporate consent management functionality into the state’s consent management platform. Initial deployment would focus on eligibility renewal monitoring, identification of services and benefits for members with medium or high-risk levels, and targeted member outreach. With a one-year implementation timeline, the AI Advocates, consent management platform, and Medi-Cal Connect would constitute the proof of concept for rollout and evaluation.

To support the success of the initial configuration, DHCS would do the following:

• Build a framework for governance, privacy, security, and regular evaluation and monitoring to identify improvement opportunities
• Define a certification process for AI Advocates
• Establish consent to share data with AI Advocates as part of the Medi-Cal enrollment and redetermination process in the California Healthcare Eligibility, Enrollment, and Retention System and in BenefitsCal
• Configure the DHCS consent management platform to manage member consent for AI Advocates
• Develop member education materials and conduct member and stakeholder outreach about the AI Advocate “app store”

After successful implementation of AI Advocates with Medi-Cal Connect, expansion may include the following:

• QHIOs. By requiring AI Advocate–compatible Qualified Health Information Organization connections in MCP contracting, DHCS could unlock consent-mediated access to member-specific, real-time clinical data, enabling incorporation of real-time clinical events into outreach workflows. AI Advocates could then respond more quickly to hospitalizations, ED visits, new diagnoses, and care transitions, strengthening follow-up and coordination for members with complex medical or behavioral health needs.
• CalSAWS. The California Statewide Automated Welfare System operates across all 58 counties as a Joint Powers Authority and supports eligibility determination and program participation for public benefits with county-supported enrollment processes. Integration with CalSAWS could enable AI Advocates to identify eligibility for additional state benefits and to detect upcoming renewals and lapses in enrollment that would contribute to churn. Links to BenefitsCal and human support at the county level would enable streamlined navigation to program enrollment.
• Additional data sources. Subsequent phases may expand access to additional data sources in a structured manner, prioritizing integrations based on feasibility, statutory authority, and measurable impact on coverage continuity. State leadership would determine the sequencing based on operational readiness. Additional data sources are described in Table 1.

Table 1. Proposed Data Sources for Future Expansions

Data Type	Data Details	Data Source	Governing Authority
CalFresh, CalWORKs, other safety-net programs	Eligibility determination, benefits calculation, benefits issuance, information management	CalSAWS	County Consortium
County behavioral health	Clinical behavioral health records	California Mental Health Services Authority (CalMHSA) semi-statewide electronic health record, with interoperability through CalMHSA Connex	CalMHSA
Housing	Housing status and location, specific health conditions, Coordinated Entry assessment	Homeless Management Information System, with five vendors	US Department of Housing and Urban Development
Immunizations	Longitudinal immunization records	California Immunization Registry	California Department of Public Health
Justice	Location, release date	County jails (system varies by county)	Counties
Employment and Disability	Quarterly wage reports, unemployment insurance claims, State Disability Insurance claims	Multiple databases	California Employment Development Department
Education (public K–12)	Enrollment, attendance, special program participation, discipline data, graduation and dropout status	California Longitudinal Pupil Achievement Data System	California Department of Education
Public utilities	Utility account status, including service address, arrears or delinquency, shutoff notices, participation in low-income rate assistance programs	Investor-owned utilities (PG&E, SCE, SDGE, SoCalGas) and publicly owned utilities (LADWP, SMUD)	California Public Utilities Commission
In-Home Supportive Services	Eligibility and case information, functional assessment data, caregiver information, authorized services and hours	Case Management, Information and Payrolling System	California Department of Social Services
Developmental disability services (Regional Centers)	Eligibility and diagnostic records, Individual Program Plans, authorizations, financial and service utilization, case management notes	Statewide Automated Needs Determination Information System	California Department of Developmental Services

Sources: CalSAWS Executive Overview (PDF), CalSAWS Joint Powers Authority, July 2024; “Electronic Health Record,” CalMHSA, accessed March 18, 2026; “CalMHSA Connex,” CalMHSA, accessed March 18, 2026; Homeless Management Information Systems: The Basics (PDF), DHCS, accessed March 18, 2026; “California Longitudinal Pupil Achievement Data System (CALPADS),” California Department of Education, accessed March 18, 2026; and “Case Management, Information and Payrolling System (CMIPS),” California Health and Human Services Agency, Office of Technology and Solutions Integration, accessed March 18, 2026.

Financing and Affordability

To enhance feasibility and affordability, this initiative would build on existing infrastructure as the primary implementation mechanism, expanding on the $315 million already invested in Medi-Cal Connect.^[xviii] California could pursue an estimated $5 million to $10 million investment, before federal matching, to develop a demonstration project, focusing on proof of concept and iterative refinement before scaling statewide. This investment includes $1 million to $2 million for technology development and $4 million to $8 million for governance infrastructure, program management, and evaluation. Estimated development costs for statewide implementation are approximately $30 million to $40 million, based on the estimated costs of building and deploying similar agentic AI tools. With the 50% federal matching, the state’s share would be approximately $15 million to $20 million, including support for ongoing leadership and stewardship at the state level.

Although the initiative requires up-front investment, cost savings are expected through improved coverage continuity and reductions in avoidable emergency department visits and hospitalizations. For members like Mei (Figure 1), early identification of crisis signals, such as missed primary care visits and medication refills, could help discover and overcome barriers in access to preventive care and build trusted linkages to lower-cost services. By shifting utilization toward prevention and by reducing churn-related administrative processing, savings would offset the initial expenditures. A portion of these savings could be reinvested to sustain operations. Additionally, the state could implement a model in which the technology vendors that develop AI Advocates for certification and inclusion pay a participation or listing fee. These vendor contributions would create a funding stream to support ongoing maintenance, oversight, and future enhancements without creating additional state budgetary pressure.

Risks and Tradeoffs

Governance and accountability. Introducing AI capabilities into public benefits programs would raise legitimate concerns about surveillance, data misuse, privacy, and security, particularly for Medi-Cal members who are affected by immigration status, employment vulnerability, or justice involvement. Community engagement, codesign, and consumer protections would be embedded in development and implementation. Trust is a precondition for adoption.

To strengthen governance and trust, DHCS would establish a detailed framework and authority for transparency, oversight, and accountability. Oversight would be conducted through existing Medi-Cal advisory bodies, including the Medi-Cal Managed Care and Mental Health Office of the Ombudsman, the Office of the Patient Advocate, the DHCS Medi-Cal Member Advisory Committee, and the Medi-Cal Voices and Vision Council, as well as the Data Exchange Framework Stakeholder Advisory Group.^[xix]

Political and stakeholder resistance may emerge from concerns about data sharing, automation, or role displacement. Delineation of responsibilities and permissions as well as demonstration of administrative relief for counties, plans, and providers would be important for sustained support.

Bias, equity, care delivery, and reach. AI models carry a risk of biased outputs, depending on the training data, design choices, and evaluation methods. Without safeguards, bias could exacerbate disparities in access to services or outreach intensity. To mitigate this risk, AI Advocates would be subject to regular bias audits, fairness testing, and ongoing monitoring, utilizing the leading local expertise at the University of California. Performance metrics would be disaggregated by race, ethnicity, geography, disability status, language, and other demographic factors to detect inequitable patterns. Identified disparities would trigger model refinement or policy adjustment.

There is also a risk that automation could be perceived as replacing human care. AI Advocates are intended to identify options to respond to rising risk levels, eligibility determination, and navigation, not to replace care management or clinical decisionmaking. DHCS could convene a governance body to develop and to implement a model for oversight and escalation to human support. This model would use research-based and industry-tested best practices to detect concerning language and to redirect members to live, person-to-person support.

AI Advocates are best suited for members who are at rising risk or who are newly eligible for services. Members with high or very high needs, including those experiencing chronic homelessness or serious mental illness, may not be effectively served through digital outreach alone. Additionally, although access to AI Advocates by phone call would be central to the design, this approach would not replace the in-person navigation that some members need or prefer. By reducing the administrative burden for lower-complexity cases, AI Advocates could enable navigators and care coordinators to dedicate their capacity to high-need members.

Data timeliness. Medi-Cal Connect primarily aggregates claims data with a lag of three to six months. This lag presents an operational barrier to the ability of AI Advocates to respond in real time to rising risk and to present timely care options to Medi-Cal members. Enabling AI Advocates to access QHIOs would alleviate this concern by incorporating real-time data, including event notifications and referrals.

Conclusion

Medi-Cal stands at a generational inflection point. California has expanded benefits and invested in whole-person care, yet federal uncertainty, fiscal pressure, workforce shortages, and administrative strain limit incremental reform. Health care costs continue to rise faster than public revenues. Coverage churn and fragmented delivery blunt the impact of recent investments. To overcome these limits, California should build on its core AI capabilities, which are now mature enough for responsible public-sector use. With clear governance and safeguards, these tools could convert existing data into earlier identification of risk and more timely connection to services.

AI Advocates would help ensure that eligible members remain covered, are identified before their needs escalate, and are connected to services when intervention is most effective. Without this shift, expanded benefits risk being unused or reaching members too late. If California does not modernize, long-term costs will climb, inequities will deepen, and confidence in the safety net will erode. California can set a bold, durable course that aligns fiscal discipline with technological maturity and that keeps Medi-Cal equitable, sustainable, effective, and responsive for all Californians.

---

Endnotes

[i] Sadiq Y. Patel et al., “Preventing Tomorrow’s High-Cost Claims: The Rising-Risk Patient Opportunity in Medicaid,” The American Journal of Managed Care 31, no. 11 (2025): 647–49, https://doi.org/10.37765/ajmc.2025.89816; Benedikt Langenberger et al., “The Application of Machine Learning to Predict High-Cost Patients: A Performance-Comparison of Different Models Using Healthcare Claims Data,” PLoS ONE 18, no. 1, e0279540 (2023), https://doi.org/10.1371/journal.pone.0279540; Luis A. Rodriguez et al., “Identifying Predictors of Homelessness Among Adults in a Large Integrated Health System in Northern California,” The Permanente Journal 27, no.1 (2023), https://doi.org/10.7812/TPP/22.096; and Eva Laura Siegel et al., “Energy Insecurity Indicators Associated with Increased Odds of Respiratory, Mental Health, and Cardiovascular Conditions,” Health Affairs 43, no. 2 (2024): 260–68, https://doi.org/10.1377/hlthaff.2023.01052.

[ii] “CalAIM: Transforming Medi-Cal,” California Department of Health Care Services (DHCS), accessed March 18, 2026; “Behavioral Health Services Act,” DHCS, accessed March 18, 2026; and “DHCS Birthing Care Pathway,” DHCS, accessed March 18, 2026.

[iii] “What Is Medi-Cal Connect,” DHCS, accessed March 18, 2026.

[iv] Kristof Stremikis and Catherine Teare, “The 25% Problem: Why Health Care Is So Expensive (And What We Can Do About It),” California Health Care Foundation (CHCF), October 13, 2025.

[v] ECM and Community Supports Quarterly Implementation Report, DHCS, updated January 2026.

[vi] Shortchanged: Health Workforce Gaps in California, CHCF, July 15, 2020; Sarah Arnquist, Addressing Medi-Cal Behavioral Health Workforce Shortages Through Non-Financial Incentives, CHCF, May 31, 2024; and California’s Direct Care Workforce Quick Reference Guide (PDF), CHCF, January 2023.

[vii] One Big Beautiful Bill Act, H.R. 1, 119th Cong. (2025).

[viii] 2022–2023 Statewide Language Survey and Implementation Plan: Report to the Governor and the Legislature (PDF), California Department of Human Resources, December 2024.

[ix] Languages Spoken by California’s Health Workforce, California Department of Health Care Access and Information, 2025.

[x] “CalAIM ASCMI Initiative,” DHCS, accessed March 18, 2026.

[xi] “What is the Model Context Protocol (MCP)?,” Model Context Protocol, accessed March 18, 2026.

[xii] “About RSST,” DHCS, accessed March 18, 2026.

[xiii] Health Insurance Portability and Accountability Act of 1996, Pub. L. No. 104–191, 110 Stat. 1936 (1996);      California Confidentiality of Medical Information Act, Cal. Civ. Code § 56 et seq. (West 2023).

[xiv] “Medicaid Administrative Claiming,” Centers for Medicare & Medicaid Services, accessed March 18, 2026.

[xv] California State Budget 2021–22 (PDF), California Governor Gavin Newsom, 2021.

[xvi] Cal. Welfare and Institutions Code § 14184.102.

[xvii] “A Healthy California for All,” California Data Exchange Framework and California Health and Human Services Agency, accessed March 18, 2026.

[xviii] California State Budget 2021–22.

[xix] “Medi-Cal Managed Care and Mental Health Office of the Ombudsman,” DHCS, accessed March 18, 2026; Compare Health Plans and Medical Groups, and Learn More About the Quality of Californians’ Health Care, Center for Data Insights and Innovation, accessed March 18, 2026; “DHCS Medi-Cal Member Advisory Committee,” DHCS, accessed March 18, 2026; and “Medi-Cal Voices and Vision Council,” DHCS, accessed March 18, 2026.