Carta Healthcare

It sells a clinical data abstraction system that combines AI software with outsourced labor. The platform ingests unstructured electronic health record data and converts it into structured registry submissions for quality reporting, reimbursement, and research. Unlike pure-play AI abstraction startups, the firm retains an in-house team of over 200 clinical abstractors who review the AI's output, which it says pushes inter-rater reliability to 98–99%.

The firm's website and available primary sources do not disclose founder names or current executive leadership. Early-stage venture backing came from Frist Cressey Ventures, whose managing partner Navid Farzad has spoken publicly in support of the company. As of the latest Altss research record, no named CEO, CIO, or board is publicly documented on the corporate site.

Hybrid Intelligence is Carta Healthcare's term for its combined AI-and-human workflow. The AI—branded Voyager for the platform layer and Lighthouse for the abstraction assistant—extracts and pre-populates registry fields from clinical documentation. Then one of the firm's 200-plus employed abstractors, all with clinical backgrounds, reviews and corrects the output. The firm argues that pure-AI abstraction alone produces errors that only experienced clinicians catch, making the blended model more reliable for payer and registry audits.

The firm's website names Memorial Hermann Health System, UNC Health, and Northern Arizona Healthcare as referenceable customers. It also references a 10-hospital system that reached three-star VQI ratings at every site after deploying the platform, though it does not identify that system by name. Carta Healthcare describes its customer base generically as trusted nationwide by hospitals and health systems.

No. The firm is a venture-backed operating company, not an investment fund, so it does not disclose assets under management or deployment figures. Total employee count outside the 200-plus abstractor pool is not public. Revenue, valuation, and profitability are undisclosed in available primary materials.