YC RFS Healthcare — Specific Problems YC Wants Solved

This page breaks down every specific healthcare problem YC has called for across its 2025 RFS editions, what the fundable solution looks like for each, and how healthcare founders should position their applications.

YC's healthcare thesis has evolved from "technology-enabled care" (where YC funded care delivery companies like Forward Health, Cerebral, and Carbon Health) to "AI that fixes healthcare's structural inefficiencies." The shift reflects two things: the maturation of AI capabilities that make previously impossible workflow automation now possible, and a growing recognition that healthcare's cost problem is primarily an administrative and coordination problem rather than a clinical one.

The specific number YC partners have cited: approximately 30% of US healthcare spending goes to administrative costs — billing, prior authorization, documentation, compliance, and coordination — not to actual care. This administrative burden is what the current YC healthcare RFS is primarily aimed at.

Clinical Documentation and the Ambient AI Scribe

The most consistently cited specific problem across multiple 2025 RFS editions: physicians spend 2+ hours per day on documentation — electronic health record entry, clinical note writing, and after-visit documentation — for every hour they spend with patients. This documentation burden is a leading cause of clinician burnout and a direct driver of reduced patient access (a physician spending 2 hours on documentation has proportionally less time for patient visits).

What YC wants: AI that listens to clinical conversations (with patient consent), understands medical context, and automatically generates accurate, structured clinical notes — reducing documentation time from hours to minutes. YC has already funded companies in this space (Abridge, Nabla, Ambience Healthcare) but considers the problem far from fully solved, particularly outside large hospital systems.

Specific gaps YC has called out:

• Ambient documentation for specialties outside primary care (emergency medicine, surgery, mental health, home care)
• Ambient documentation that integrates natively with all major EHR platforms rather than requiring a parallel workflow
• Documentation AI for non-English-speaking patient populations

Prior Authorization Automation

Prior authorization — the process by which insurers require clinicians to request approval before ordering specific tests, medications, or procedures — is described by YC as one of the most wasteful and clinically harmful administrative processes in American healthcare. The numbers: physicians report spending an average of 13 hours per week on prior authorization; 94% of physicians report that prior auth delays or denials have harmed patients.

What YC wants: AI that automates the prior authorization process end-to-end — pulling relevant clinical evidence from the patient record, matching it to the insurer's specific coverage criteria, submitting the request, and tracking the response — without requiring physician time for routine authorizations.

Specific gaps YC has called out:

• Real-time prior auth (pre-service determination before the patient encounter) rather than retrospective submission
• Prior auth AI that works across all payer types (commercial, Medicare Advantage, Medicaid) and their different criteria systems
• Appeals automation for denied authorizations

Medical Coding and Revenue Cycle Automation

Translating clinical encounters into the billing codes (ICD-10, CPT, HCPCS) that insurers use for reimbursement is a skilled, time-intensive task performed by specialized medical coders. Coding errors cost healthcare providers an estimated $125B per year in denied or underpaid claims, while over-coding creates compliance exposure.

What YC wants: AI that reads clinical documentation and accurately generates complete, defensible medical codes — reducing the need for large medical coding teams and improving the accuracy and completeness of coding without increasing compliance risk.

Specific gaps:

• Coding AI for complex specialties (surgery, oncology, cardiology) where code selection requires deep clinical judgment
• Audit trail and explainability for AI-generated codes that satisfies compliance requirements
• Real-time coding feedback for clinicians during documentation (rather than retrospective correction)

Healthcare Revenue Cycle Management

Beyond coding, the broader revenue cycle — claims submission, denial management, payment posting, patient billing, and collections — remains heavily manual at most health systems. Denial rates for initial claims run 5-10% at many providers, and working those denials requires significant administrative staff.

What YC wants: AI that automates the full revenue cycle workflow — from claims scrubbing before submission through denial appeal to final payment reconciliation — reducing staff requirements and improving net collection rates.

Patient Communication and Engagement

The patient experience outside clinical encounters — appointment scheduling, care instructions, medication reminders, test result communication, and follow-up coordination — is largely handled through systems designed for manual operation (phone trees, generic patient portals) that produce poor patient engagement and significant staff burden.

What YC wants: AI-powered patient communication that is conversational, personalized, and proactive — reducing the administrative burden on front desk and care coordination staff while improving patient adherence and satisfaction. This category has overlap with the Voice AI RFS category: YC specifically wants voice AI that can handle high-volume patient communication workflows (appointment reminders, pre-visit instructions, post-discharge follow-up) without human involvement for routine interactions.

Diagnostic AI for High-Volume, High-Error-Rate Conditions

YC has called for AI that improves diagnostic accuracy in conditions where diagnostic error rates are high, the patient population is large, and the consequences of missed diagnosis are severe. Specific conditions that meet these criteria and have been discussed in YC partner commentary:

• Sepsis early warning: Sepsis is the leading cause of hospital deaths and is frequently missed until late stages; AI pattern recognition across vital signs and lab values can identify patients hours before clinical deterioration
• Diabetic retinopathy screening: A preventable cause of blindness with a large at-risk population (all diabetic patients) and a screening workflow that is scalable with AI image analysis
• Skin condition triage: First-line dermatology screening at primary care or consumer level
• Mental health screening: Scalable identification of patients at risk for depression, anxiety, or suicidal ideation through structured screening in primary care settings

The pattern: companies addressing the administrative burden (documentation, prior auth, revenue cycle) raised larger rounds faster than companies addressing clinical decision support or diagnostic AI, reflecting the more immediate ROI and shorter enterprise sales cycle for administrative automation.

Specificity of the workflow problem

The most fundable healthcare applications in YC's portfolio describe a specific workflow failure — not "healthcare is inefficient" but "emergency physicians spend 47 minutes per shift on documentation that our product reduces to 8 minutes." The more specific the workflow, the user, and the measurable outcome, the more credible the application.

Evidence of clinical validity

Healthcare enterprise buyers require evidence that AI tools are accurate and safe before deploying them clinically. The most fundable early-stage healthcare applications have either a pilot with a health system that has produced outcome data, or a study design that will produce outcome data within a defined timeline. Applications that describe the AI product without addressing accuracy or clinical validation leave a critical question unanswered.

Regulatory pathway clarity

Many healthcare AI products require FDA clearance (particularly diagnostic AI) or at minimum a legal analysis of whether they qualify as medical devices. YC-funded healthcare companies address the regulatory pathway explicitly — either confirming that their product does not require FDA clearance and explaining why, or describing their 510(k) or De Novo pathway if it does.

HIPAA and enterprise security posture

Healthcare enterprise buyers will not pilot any product that cannot demonstrate HIPAA compliance and a credible security posture. The most fundable healthcare applications address this directly in the application or describe a concrete plan to achieve SOC 2 Type II and BAA signing capability before enterprise pilots begin.