Merck co-leads Faro AI's $37.3M bet on structured clinical-trial agents

Why it matters

Faro is betting that clinical AI will be won below the chatbot layer. Merck's co-lead investment gives Scott Chetham capital and pharma-side validation to turn structured protocols into a control system for regulated agents.

Scott Chetham's Faro AI raised a $37.3 million Series B to extend its structured clinical-development platform into AI agents that can work across study design, execution and regulatory workflows. Merck Global Health Innovation Fund and S32 co-led the financing, according to Faro's August 26 announcement and a report by FinSMEs.

General Catalyst, Northpond Ventures, Polaris Partners, PTX Capital and Zetta returned for the round. Ankona Capital joined as a new investor. Faro did not disclose a valuation.

Chetham co-founded Faro with physician and venture investor Ross Jaffe in 2019 after years spent close to the operational machinery of clinical research. Before Faro, Chetham led clinical research operations and data management at Verily Life Sciences, formerly Google Life Sciences. General Catalyst said in its 2023 investment note that he had managed the development of more than 100 clinical-trial protocols across 54 programs with biopharma sponsors.

His earlier work included clinical roles at ImpediMed, serving as CTO and vice president of clinical operations at Intersection Medical, and a stint as a venture partner at Versant Ventures. Jaffe co-founded Versant and trained as a physician, completing an internal-medicine residency at UCSF after earning an M.D. from Johns Hopkins and an M.B.A. from Stanford.

That mix of clinical operations, product development and venture investing shaped Faro's founding thesis. Chetham had watched trial timelines slip while protocols remained long, static documents, disconnected from the data and operational choices they controlled. Faro began with the premise that the protocol should be structured and machine-readable from its first draft.

The protocol becomes the agent's memory

Faro's original product helped clinical-development groups design studies, compare protocol choices and generate documents. Chetham is now pushing the underlying data model further downstream.

Clinical protocols contain linked scientific, medical, regulatory and operational decisions. Changing an endpoint can alter eligibility rules, site procedures, data collection, statistical analysis and submission materials. A general-purpose language model can draft text describing those changes. An agent expected to configure systems or flag contradictions needs a persistent representation of how the decisions connect.

Faro says its proprietary ontology converts those relationships and constraints into structured intent that software can reason over. The Faro platform currently covers study design, document authoring and workflow automation, including automated electronic data capture builds. The Series B will fund agents intended to move across a wider portion of the development process while preserving the original clinical intent and human oversight.

Faro's financing materials frame the structured data layer as the foundation for its agent strategy. The distinction matters because document generation has become an increasingly common feature across enterprise AI products. A model that produces a plausible protocol section is useful. A system that understands how the section changes the rest of a regulated study has a better chance of becoming infrastructure.

Faro's work also aligns with emerging clinical-data standards. The Clinical Data Interchange Standards Consortium says Faro Study Designer authors protocols in a format aligned with the Unified Study Definitions Model, connecting biomedical concepts, controlled terminology and operational information inside one data model. Standards give Faro a route for moving protocol data between tools rather than trapping it inside another proprietary document layer.

Merck brings a customer-side view of the problem

Merck Global Health Innovation Fund's role as co-lead carries more weight than a passive strategic investment. Merck sits inside the kind of large, regulated organization where Faro's agents will have to prove they can survive procurement reviews, data-governance controls and clinical oversight.

Faro's homepage cites a 2024 peer-reviewed paper by Chetham and other researchers that applied a method for redesigning schedules of assessment to Merck protocols. The work examined potential effects on patient and site burden, costs and trial complexity. That history does not establish the scope of any current commercial relationship, though it shows Faro's structured-protocol approach has already been applied to Merck study materials.

Faro says six of the world's 10 largest pharmaceutical manufacturers use its platform. Faro has not named all six in the financing materials. The San Diego developer also reports identifying more than $300 million in potential cost savings, 200,000 potential patient hours avoided and more than 1,500 potential site-staff RVUs avoided. Those figures describe modeled or potential impact reported by Faro, rather than audited savings already captured by customers.

The language is important. Clinical-trial software vendors can estimate labor avoided or costs removed from a proposed protocol long before a sponsor completes the study and realizes the financial benefit. Faro's Series B investors are backing the data foundation and its adoption inside major drugmakers, rather than a disclosed revenue or savings figure.

A crowded market is converging on structured protocols

Faro has competitors pursuing the same shift away from document-bound clinical development. Nurocor offers structured protocol elements that can be reused across study documents, with changes cascading from a single source. AstraZeneca's Evinova combines a connected USDM data architecture with study design, costing, feasibility and protocol digitization. Carelane approaches the problem from trial execution, using agents to convert protocols into electronic case-report forms and other study infrastructure.

Chetham's wager is that Faro's ontology can serve as the control layer beneath these workflows. The product advantage will depend on how much clinical intent Faro can encode, how well that intent transfers into customers' existing systems, and whether agents can operate reliably as protocols change.

The round gives Chetham substantially more capital for that test. Faro had raised $18.3 million in total after its $15 million Series A in 2021, then added $20 million led by General Catalyst in 2023. Adding those disclosed amounts to the Series B puts Faro's publicly announced funding at least at $75.6 million.

Seven years after Chetham set out to replace the static protocol, the enterprise AI market has arrived at the same underlying problem from the opposite direction. Agents need structured context before pharmaceutical developers can trust them with connected, regulated work. Faro spent years building that context. The Series B will show whether Chetham can turn it into a durable operating layer across clinical development.