CMS_RFI-Sources Sought_AI Tools for Medicare.gov (002).pdf

AI Tools for Medicare Experience Modernization - Request for Information

Notice Type: Sources Sought

Notice ID: 269999

Published Date:

Response Date: March 31, 2026 (Responses will be considered as received)

Description

This is a Request for Information (RFI) only. This is NOT a solicitation for proposals, proposed abstracts, or quotations, and it does not commit the Centers for Medicare and Medicaid Services (CMS) to award a contract now or in the future. Likewise, responding to this RFI does not guarantee participation in or award under any future procurement.

The Centers for Medicare & Medicaid Services (CMS) seeks information regarding artificial intelligence (AI) and machine learning (ML) platforms and services that are capable of enhancing the Medicare experience for beneficiaries across digital and voice channels.

CMS is interested in solutions that can provide personalized plan recommendations, conversational AI support, predictive analytics, accessible decision support tools, and call center automation to help Medicare beneficiaries make informed plan choices and maximize their coverage.

Background

Approximately 70 million Medicare beneficiaries navigate coverage options across Medicare.gov, Medicare Plan Finder, and the 1-800-MEDICARE Call Center. These beneficiaries often must choose from multiple health plan types —including Medicare Advantage, Medicare Supplement (Medigap), Part D Prescription Drug Plans, and Special Needs Plans —each with varying benefits, provider networks, formularies, and coverage rules. Current tools rely primarily on static comparison tables and documents that can be difficult to navigate, particularly for beneficiaries with limited health literacy, language barriers, or cognitive challenges. During peak enrollment periods, wait times at the call center may be longer as call center representatives must explain the varying amount of coverage options and health plan types orally over the phone.

CMS seeks to leverage emerging AI technologies to achieve the following objectives:

• Provide personalized, claims-informed plan recommendations based on individual health needs, medications, provider preferences, and actual utilization patterns

• Offer real-time conversational AI support through chatbots, virtual assistants, and AI voice advisors that are available 24/7

• Use predictive analytics to match beneficiaries with plans that are likely to meet their needs

• Translate complex plan information and Medicare documents into accessible, plain-language explanations

• Ensure equitable access to decision support across diverse populations

• Automate routine call center inquiries while seamlessly escalating complex cases to human Customer Service Representatives (CSRs)

• Enable 100% call quality assurance through AI-powered call analysis and sentiment detection

• Generate insights for CMS about beneficiary preferences, plan performance, and operational improvements

• Create seamless experiences across Medicare.gov, Plan Finder, and 1-800- MEDICARE with context preservation between channels

Purpose of this RFI

CMS is issuing this RFI to accomplish the following:

• Identify vendors with production-ready, Medicare-specific AI/ML capabilities for healthcare decision support and call center automation

• Understand the current state of commercial AI solutions for Medicare plan selection and beneficiary support

• Gather information about AI approaches to plan recommendation, conversational support, predictive analytics, and call center intelligence

• Learn about privacy-preserving AI techniques that are suitable for sensitive beneficiary data

• Understand implementation approaches, timelines, and cost structures

• Identify organizational qualifications, market readiness, and volume thresholds

• Assess vendor independence from insurance sales entities to ensure plan neutrality

• Inform the development of future solicitations for AI Tools for Medicare Experience Modernization

Market Readiness & Production Status Requirements

Active Production Deployment

Respondents must demonstrate that their AI solutions meet the following criteria:

• The solutions are currently active and live in production environments, not aspirational, pilot-only, or in development

• The solutions are serving real beneficiaries or customers with measurable outcomes and performance metrics

• The solutions are proven at scale with documented uptime, reliability, and performance benchmarks

Medicare-Specific Experience & Volume Thresholds

For AI Plan Finder / Coverage Guidance Solutions:

Solutions must demonstrate the following:

• At least 1 million Medicare enrollments annually processed through their AI-powered systems across a variety of plan types and carriers representative of the general Medicare population

• Multi-carrier and multi-plan type experience across Medicare Advantage, Medicare Supplement (Medigap), Part D Prescription Drug Plans, and Special Needs Plans (C- SNPs, I-SNPs, D-SNPs)

• Medicare-specific knowledge and data integration, including formularies, provider networks, benefit structures, and CMS compliance and regulatory requirements

• Proven accuracy in producing tailored plan recommendations, benefit explanations, and eligibility determinations

For Call Center AI Solutions:

Solutions must demonstrate the following:

• Active deployment in healthcare or government call center environments with documented performance improvements

• High-volume call handling experience with a minimum of 250,000 calls or equivalent minutes annually for voice automation and for QA/intelligence platforms

• Medicare-specific implementations are preferred, but general healthcare or government call center AI tools will be considered if capabilities are transferable

Organizational Independence & Conflict of Interest Requirements

Prohibited Affiliations

To ensure plan neutrality and beneficiary-first guidance, respondents must not be affiliated with or owned by any of the following:

• Insurance carriers or health plans that offer Medicare products

• Insurance producing agencies or brokerages that earn commissions from Medicare plan sales

• Field Marketing Organizations (FMOs) or agencies of any kind whose business model relies on insurance sales commissions

• Any entity with a financial incentive to steer beneficiaries toward specific plans or carriers

Proprietary Technology & Data Ownership

Respondents must demonstrate the following:

• Ownership of their AI technology, including algorithms, models, and intellectual property

• Proprietary data assets or licensed data with full resale and integration rights

• Control over AI training, deployment, and governance without dependencies on third-party AI platforms that could compromise CMS data security or beneficiary privacy

Exception for Third-Party Data: Organizations that leverage third-party data sources (e.g., plan benefit data, provider directories) may be considered if they meet the following conditions:

• They possess full resale and integration rights for the data

• Third-party data vendors agree to comply with CMS security, privacy, and HIPAA requirements

• The respondent maintains control over AI logic and beneficiary interactions

Information Requested

Respondents should address the following topic areas:

1. Organizational Overview & Market Readiness

In this section, please include:

• Company profile, ownership structure, and independence attestation

• Production deployment evidence, including volume metrics, client references, and performance data

• Medicare-specific versus general AI capabilities

• Financial stability and organizational capacity

• Years in operation and relevant industry experience

2. AI/ML Platform Capabilities

Respondents should address:

• Your AI/ML platform and its core capabilities

• Your experience with Medicare plan selection, coverage guidance, or beneficiary decision support systems

• Natural language processing and conversational AI capabilities

• Predictive analytics and machine learning model types

• Model explainability and transparency features

• Bias detection and fairness testing capabilities

• Call center AI capabilities, including voice automation, QA/intelligence platforms, and CSR assist tools

3. Medicare and Government Experience

This section should cover:

• Your experience working with Medicare, Medicaid, or other government health programs

• Annual Medicare enrollment volume processed through AI-powered systems

• Breakdown by plan type (Medicare Advantage, Medigap, Part D, SNPs) and carrier diversity

• Your experience with health plan comparison or beneficiary decision support tools

• Your understanding of managed care plan structures, benefits, and Medicare regulatory requirements

• Your familiarity with health insurance literacy challenges

• Previous work with vulnerable or diverse populations

• Call center experience in healthcare or government environments (if applicable)

4. Technical Architecture and Integration

Please detail:

• Cloud deployment capabilities (AWS preferred)

• Integration approaches with CMS data assets, including claims, enrollment, eligibility, and provider directories

• API design and interoperability standards supported (FHIR, X12, RESTful)

• Bidirectional intelligence layer for cross-channel connectivity across Medicare.gov, Plan Finder, and 1-800-MEDICARE

• Session context preservation and handoff capabilities between digital and voice channels

• Data security and privacy architecture

• Scalability to support millions of beneficiaries during peak enrollment periods

• Performance characteristics, including response times, concurrent users, and uptime SLAs

5. Privacy, Security, and Compliance

Respondents should explain:

• Your HIPAA compliance approach and audit history

• Privacy-preserving AI and machine learning techniques

• FedRAMP authorization status or path to authorization

• Your experience with FISMA Moderate environments

• Data governance and retention practices

• Your approach to handling Protected Health Information (PHI) and Personally Identifiable Information (PII)

• Data encryption standards (at rest and in transit)

• Incident response and breach notification protocols

6. Accessibility and Usability

In your response, address:

• Your Section 508 / WCAG 2.1 AA compliance approach and audit results

• Support for assistive technologies

• Multilingual capabilities and approach to translation accuracy

• Plain language and health literacy features

• Adaptive interfaces for varying literacy levels and cognitive abilities

• Mobile and responsive design capabilities

• User experience design philosophy and beneficiary testing evidence

7. Model Development and Maintenance

Please outline:

• Your approach to developing and training recommendation models

• Data requirements for model training

• Model validation and testing methodology

• Your approach to continuous model improvement and retraining

• Model monitoring and performance tracking

• Your handling of model drift and degradation

• Continuous learning mechanisms based on beneficiary feedback and outcomes

8. Fairness, Bias, and Equity

Describe your approach to:

• Testing AI models for bias

• Ensuring equitable outcomes across demographic groups

• Addressing health disparities through AI

• Fairness metrics and monitoring approaches

• Remediation strategies when bias is detected

• Transparency in AI training data sources and model governance

9. Claims-Informed Personalization & Data Integration

This section should include:

• Your approach to leveraging Medicare claims data, including Part A&B, Part D, and MA encounters

• Auto-population capabilities for logged-in beneficiaries, including prescriptions, providers, and utilization history

• Beneficiary consent mechanisms for claims-informed personalization

• Privacy-enhancing technologies for AI model training with claims data

• Provider relationship identification and continuity of care alignment

• Medication adherence improvement through formulary matching

10. Call Center Capabilities (if applicable)

Please discuss:

• Voice automation use cases and performance metrics, including automation rate and handle time reduction

• AI call intelligence platforms, including 100% QA coverage, sentiment analysis, and compliance monitoring

• CSR assist tools that provide instant decision support tools, personalized and contextual via CMS data APIs

• Agentic voice advisor capabilities with seamless escalation to human CSRs

• Post-call digital follow-ups tied to call center interactions

• Correlation of digital friction with call volume drivers

11. Implementation Approach

Respondents should detail:

• Your proposed implementation methodology and timeline

• Phased rollout strategy from pilot to national scale

• Training and change management approach for CSRs, SHIP counselors, and enrollment assisters

• Required data and integration points

• Dependencies and assumptions

• Risk mitigation strategies

12. Cost and Pricing Structure

Responses to this RFI are not offers and cannot be accepted by the Government to form a binding contract. All pricing information requested below should be provided as a Rough Order of Magnitude (ROM) estimate only and is non-binding. These estimates are for informational and planning purposes only and will not be used for source selection or contract award.

In this section, explain:

• Your pricing model (per beneficiary, per transaction, subscription, etc.)

• Cost drivers and variables

• Typical implementation costs

• Ongoing operational costs

• Total cost of ownership over 5 years

• Pricing for proof-of-concept versus production deployment

13. References and Past Performance

Please provide:

• Relevant projects with similar scope, scale, or complexity

• Customer references for AI/ML Medicare or healthcare implementations

• Demonstrated outcomes and impact metrics from previous work

• Case studies showing beneficiary satisfaction improvements, coverage matching accuracy, or operational efficiency gains

RFI Response Instructions

Response Deadline: March 31, 2026 (Responses will be considered as received)

Submission Method: PDF Format via email to aaron.blackshire@cms.hhs.gov and matthew.wendel@cms.hhs.gov.

Page Limit: Narrative responses within the form should be concise; supplemental attachments (if permitted in form) should not exceed 20 pages total

Questions: All questions regarding this RFI should be submitted via email to:

Primary Contact: Matthew Wendel, USDS Supporting CMS Email: matthew.wendel@cms.hhs.gov

Alternate Contact: Aaron Blackshire, Contracting Officer Email: Aaron.blackshire@cms.hhs.gov

Important Notes

Please note the following important information:

• This is an RFI only and does not constitute a commitment by CMS to issue a solicitation or award a contract.

• Responses to this RFI will be used to inform market research and development of potential future solicitation requirements.

• CMS will not reimburse respondents for any costs associated with preparing RFI responses.

• Proprietary or confidential information should be clearly marked.

• CMS may conduct follow-up discussions with selected respondents.

• Information provided in response to this RFI may be shared with other government agencies.

• A future solicitation, if issued, may differ substantially from the concepts described in this RFI.

mailto:aaron.blackshire@cms.hhs.gov

• CMS reserves the right to use information gathered through this RFI in its evaluation of future proposals.

Anticipated Procurement Approach

Following this RFI and subject to the availability of funds and agency priorities, CMS anticipates issuing formal solicitations for AI Tools for Medicare Experience Modernization.

Various procurement methodologies may be employed, potentially including the following:

• Challenge-based acquisition approaches with proof-of-concept evaluations

• Multiple awards for initial demonstration phases

• Head-to-head evaluation of AI capabilities against defined scenarios

• Down-selection to vendors for production implementation

• Phased rollout starting with pilot states

• Aggressive base-period affordability objectives for proof-of-concept phases

Proof-of-Concept Assessment Metrics (Illustrative)

If CMS proceeds with a challenge-based acquisition, contractor solutions may be evaluated against metrics such as the following:

• Plan Recommendation Accuracy: ≥85% alignment with expert recommendations across test scenarios; explainability provided for all recommendations

• Conversational AI Quality: ≥80% user satisfaction in test dialogues; successful resolution of ≥75% of common questions; appropriate escalation of complex queries

• Response Time: Recommendation engine p95 ≤2s; Chatbot p95 ≤3s

• Accessibility Compliance: WCAG 2.1 AA conformance verified by third-party;

successful use by testers with disabilities

• Multilingual Support: Accurate support for at minimum English and Spanish;

additional languages proposed

• Privacy and Security: HIPAA compliance verified; no PII exposure; successful privacy impact assessment; security controls documented

• Fairness and Bias: Documented bias testing; no statistically significant disparate impact on protected classes; mitigation strategies for identified bias

• User Experience: ≥75% usability testing success rate; System Usability Scale (SUS) score ≥70

• Integration Capability: Successful integration with test data feeds; documented APIs; data quality validation

• Explainability: Clear explanations provided for all recommendations; users can understand basis for suggestions

Production Implementation Requirements (Illustrative)

If CMS proceeds to production implementation, requirements may include the following:

Data Integration & Infrastructure:

• Bidirectional intelligence layer connecting Medicare.gov, Plan Finder, and 1-800- MEDICARE Call Center

• Real-time API access to CMS data assets, including Part A&B claims, Part D claims, MA encounters, enrollment/eligibility, provider directories, and call recordings

• Session context preservation and handoff between digital and voice channels

AI Governance & Responsible AI Framework:

• Explainability for all AI-generated recommendations

• Continuous validation against authoritative CMS source data

• Strict controls to prevent misleading responses

• Modular architecture enabling safe iteration

• Beneficiary consent mechanisms for claims-informed personalization

• Plan neutrality preservation

Call Center-Specific Requirements:

• 100% call recording analysis with automated QA scoring

• AI call intelligence for sentiment analysis and operational insights

• CSR assist tools providing instant personalized, contextual decision support tools

• Voice automation for targeted use cases

• Agentic voice advisor with seamless escalation to human CSRs

Cross-Channel Experience Requirements:

• Differentiation between logged-in and logged-out experiences

• Auto-population of beneficiary data for authenticated users

• Post-call digital follow-ups

• Correlation of digital friction with call volume drivers

Performance & Operational Metrics:

• Measurable improvements in coverage matching accuracy, beneficiary satisfaction, digital engagement, call volume reduction, QA cost savings, and peak-period scalability

• Continuous learning mechanisms with model retraining

Security & Compliance:

• HIPAA compliance with privacy-enhancing technologies

• Read-only claims data access with beneficiary consent workflows

• Audit trails for all AI-generated guidance

• FedRAMP authorization for cloud services

Classification Information

Department/Agency: Department of Health and Human Services / Centers for Medicare and Medicaid Services

Office: Office of Acquisition and Grants Management

NAICS Code: 541512 - Computer Systems Design Services

PSC Code: D399 - IT and Telecom - Other IT and Telecommunications

Set-Aside: No set-aside anticipated (full and open competition)

Place of Performance: Contractor site; some meetings at CMS facilities in Baltimore, MD

Contact Information

Primary Point of Contact:

Matthew Wendel, USDS Supporting CMS Email: matthew.wendel@cms.hhs.gov

Alternative Point of Contact:

Aaron Blackshire, CMS Office of Acquisition and Grants Management Email: aaron.blackshire@cms.hhs.gov

Contracting Office Address:

Centers for Medicare & Medicaid Services Office of Acquisition and Grants Management 7500 Security Blvd Baltimore, MD 21244

USA

Disclaimer

Responses to this RFI are voluntary. Failure to respond to this RFI will not preclude participation in any future procurement.