Exclusive: Ex-Palantir execs raise $12 million for AI drug discovery startup Perceptic

“For years, the industry has tried to improve each part of the [drug discovery] process separately, but that’s a linear process where insight dies at every handoff,” Flock said. 

Perceptic’s platform is “infrastructure and model agnostic,” Flock said, meaning customers can plug in their own data, hardware and AI models, while Perceptic acts as the layer that ties them together.

Perceptic is targeting three areas of pharma R&D. The first is scouting external assets that biotechnology companies have developed and that big pharma companies look to license. The startup says its system can compress the scientific due diligence needed to assess these drug candidates from weeks to hours.

The second area where Perceptic is focusing is helping pharma companies choose which indications to pursue in clinical trials, a decision that Flock said can swing the fate of investments worth millions of dollars.

The third is building a “data foundation” for clinical trial design which the company says has produced a 50-fold increase in clinical data extractions.

Sonali De Rycker, the Accel partner who led the firm’s investment, said she was won over by the idea that Perceptic’s software can “follow the drug” through the entire life cycle of development rather than being pitched towards a particular departmental silo within a big pharma company. “From the point at which you have hypothesis and evidence all the way to when you’re designing the clinical trial, and everything you do in between … it makes no sense for it to be siloed,” she told Fortune.

De Rycker said Accel had been tracking Flock and his co-founders, Martin Copes and Zaki Trache, while they were still at Palantir, where the trio were key engineers on AIP, and met them shortly after they decided to leave. The firm invested roughly a year after that first meeting, by which point Perceptic had moved beyond pilots into paid production deployments, she said. The team has grown to about 20 employees today, Flock said.

Pharmaceutical companies, Flock said, tend to draw on three buckets of data: public knowledge, such as patents and literature; internal proprietary data accumulated over years of research and clinical trials; and external datasets purchased from consultants and data vendors. Perceptic can harmonize all three types of data, he said. The system uses “AI workers,” or AI agents, that are tuned to different data types to hunt for insights or optimizations. 

Pharmaceutical customers need to know the provenance of any data used to make a decision, Flock said. As a result, they cannot tolerate AI hallucinations, where an AI model invents or conflates information. Perceptic’s AI system allows customers to trace every claim back to its source, he said.

De Rycker argued that Perceptic’s approach reflects a broader pattern in enterprise AI, in which platforms increasingly unify workflows and data across multiple departments rather than offering standalone tools. She added that these platforms can be “almost a new source of truth,” potentially replacing—or at least relegating to the background—traditional databases and enterprise resource planning software. 

She added that Perceptic has a “right to win” from Europe given the concentration of pharmaceutical talent in Switzerland and the U.K. Most of the company’s engineering is in London, drawing in part on Palantir alumni, Flock said; many of its customers are in the United States, where Perceptic plans to expand its presence.

Nathan Benaich, the founder and general partner of Air Street Capital, said in a statement that pharma’s next R&D leap “won’t come from a thousand better point tools, or from frontier models alone” but from an operating system “that connects data, decisions, and context across a 15-year process.”

Flock said the bulk of the seed money will go into engineering and growing Perceptic’s customer base. “We’re far beyond product-market fit,” he said. “It’s about scaling out.”