Qiagen N.V. is betting that drug discovery agents need a trustworthy knowledge foundation as much as capable models. For biopharma companies, that means giving agents information with clear provenance and enough context to support their answers.
Knowledge graphs can provide that context layer for AI agents, according to Iman Bhattacharya, senior global product marketing manager at Qiagen. However, human curation remains central to the approach.
“I think while everybody is talking about AI and the model layer, we deeply focus on the knowledge foundation and the knowledge content of it,” he said. “And we believe that AI is here to stay. AI is here to elevate yourself, but it has to be deeply grounded in human curation.”
Bhattacharya spoke with theCUBE Research’s John Furrier at GraphSummit, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how curated knowledge and traceable data can ground AI agents used in drug discovery. (* Disclosure below.)
Drug discovery agents draw on curated knowledge
The company’s bioinformatics arm has manually curated biomedical data for more than 25 years, relying on more than 150 MD- and PhD-level experts, Bhattacharya noted. That work now anchors the Qiagen Discovery Platform, which layers Model Context Protocol, or MCP, access and an agentic Discovery Explorer on top of the curated knowledge base. The company also teamed with Nvidia Corp. in May to advance graph-based AI for drug discovery.
“It’s not just data or it’s not just knowledge foundation. It’s actually a complete system,” Bhattacharya said. “The goal is not to make it siloed. The goal is to make it be a part of that ecosystem.”
Accuracy is where that design matters most. Pharma customers judge return on investment by how fast they can reach accurate indications ahead of rivals, and agents that invent answers undercut that goal, according to Bhattacharya.
“Your agent will start to provide output, and it will hallucinate. It will never say no. We have seen that it will provide synthetic results,” he said. “But ultimately, what matters is you produce something, your output is something which is good and traceable.”
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of GraphSummit:
(* Disclosure: TheCUBE is a paid media partner for GraphSummit. Neither Neo4j, the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)