What You Should Know
- Applied artificial intelligence and healthcare navigation company Baba secures $40M in Series A financing led by FirstMark Capital with participation from CVS Health Ventures and value-based healthcare specialist Valtruis, alongside returning venture investors General Catalyst, Genius Ventures, and Ground Up Ventures.
- Introduces BabaBench (published on arXiv and accepted to an upcoming NeurIPS workshop): a specialized AI benchmark designed to evaluate how frontier autonomous agents execute long-horizon patient advocacy workflows (multi-week insurance phone calls, complex billing appeals, prior authorization disputes) from start to finish while enforcing strict privacy safeguards.
Redefining the Unit Economics of Complex Care
Historically, personalized patient advocacy has been an elite concierge product, accessible only to affluent families capable of paying out-of-pocket for private geriatric care managers. Traditional Medicare fee-for-service payment rails provide virtually no dependable margin for high-touch navigation, driving up the cost of managing complex, multimorbid patients to unsustainable levels.
Baba’s thesis centers on leveraging applied artificial intelligence to compress the administrative overhead of care coordination, making continuous longitudinal advocacy economically viable across publicly insured populations.
Under Baba’s hybrid care delivery model, every enrolled patient is paired with a dedicated human care team that includes care coordinators and licensed clinical social workers or therapists. The human staff is backed by proprietary AI agents that manage the logistical heavy lifting: synchronizing specialist referrals, tracking hospital-to-home discharge transitions, coordinating in-home aide support, troubleshooting billing denials, and securing community resources such as housing assistance and medical transportation.
To date, Baba has supported more than 10,000 patients and families nationwide, with more than 100,000 families currently on its waitlist, reflecting massive latent demand for patient-facing logistical navigation.
BabaBench: Testing Frontier Agents Against Long-Horizon Execution
Alongside its clinical operations, Baba is carving out a technical niche within applied AI research with the release of BabaBench, a benchmark specifically designed to evaluate how frontier foundation models and autonomous agents execute complex, long-horizon healthcare tasks.
Most standard AI evaluations test conversational fluency, factual retrieval, or isolated clinical board exam questions. In contrast, BabaBench evaluates whether an agent can autonomously complete an end-to-end bureaucratic workflow across weeks or months—such as disputing an erroneous balance bill, parsing insurer coverage criteria, or completing multi-page social program applications—without generating prohibited clinical actions or triggering protected health information (PHI) privacy violations.
The benchmark, released as a preprint on arXiv, has been accepted for presentation at an upcoming workshop at the Conference on Neural Information Processing Systems (NeurIPS), establishing a standardized framework for measuring autonomous execution in highly regulated, high-stakes healthcare environments.
Federal Policy Alignment: Scaling Through the CMS ACCESS Model
Baba’s expansion coincides with a decisive shift within the Centers for Medicare & Medicaid Services (CMS) away from transactional fee-for-service billing and toward technology-supported chronic disease management.
In June, the CMS Innovation Center (CMMI) selected Baba as an operating participant in the ACCESS Model (Advancing Chronic Care with Effective, Scalable Solutions), a multi-year demonstration testing outcomes-based reimbursement for technology-enabled chronic care. Notably, Baba secured qualification across all four clinical tracks under ACCESS:
- Early Cardio-Kidney-Metabolic (CKM)
- Advanced Cardio-Kidney-Metabolic
- Musculoskeletal Health
- Behavioral Health
Participation in ACCESS positions Baba to bypass fee-for-service constraints, tying its revenue model directly to verified patient outcomes and cost reductions. The company has complemented this federal integration with cross-sector partnerships spanning national commercial and Medicare Advantage health plans, academic health systems, and condition-specific advocacy organizations, including recent collaborations with the American Stroke Foundation and the American Breast Cancer Foundation.
To manage its scaling clinical footprint, Baba recently bolstered its senior leadership bench, recruiting Mount Sinai veteran Alicia Tennenbaum, LCSW, as Executive Director of Clinical Operations, alongside product designer Alan Li, formerly of financial software unicorn Ramp.
“By dramatically lowering the cost of operating and coordinating care, we have an opportunity to make care delivery economically viable for populations that have historically been too expensive to serve,” said Connor Sweeney, Founder and CEO of Baba. “At Baba, we bring together providers, care coordinators, licensed therapists and AI, supporting people across their medical, financial, social and behavioral health needs. Causal, dollar-denominated care navigation has been a holy grail in healthcare for a long time, and it’s finally possible because of AI.”
