Digital Science’s Sebastian Schmidt discussed how enterprise AI can connect scientific discovery with traceable evidence ahead of AI & Big Data Expo Europe in Amsterdam.
Sebastian, who is the EVP of Enterprise at Digital Science , outlines the company’s approach to linking proprietary research with publications, patents, and clinical trial records. He describes a semantic layer that uses common ontologies and linked data structures to connect internal and external sources, with outputs traceable to their underlying records.
Our Q&A with Sebastian covers human expert sign-off, literature review workflows, patent intelligence, and the role of source provenance in scientific and legal decisions. Sebastian sets out the mechanisms and benefits he attributes to these approaches, rather than presenting measured results from a named enterprise deployment.
Sebastian Schmidt: Consumer software tolerates text generation errors because the stakes are low, but enterprise systems should treat generative AI primarily as a hypothesis engine. To protect against financial and legal liability, enterprise systems must ground AI outputs using reference data and structured ontologies.
Rather than replacing human judgement, these models then function as discovery catalysts—identifying promising pathways while mandating that every chemical formula, target interaction, or experimental parameter is validated against structured domain ontologies.
By grounding generative outputs in established scientific evidence, research divisions can innovate rapidly while ensuring that no unverified formula enters a commercial pipeline, regulatory filing, or legal process without direct human expert sign-off and proven backing.
Sebastian: Standard web crawls capture noisy, unverified internet chatter, whereas pharmaceutical discovery demands rigorous, evidence-based data. Enterprise AI requires peer-reviewed research grounded in two core pillars: trusted reference data and structured ontologies.
Trusted reference data provides the precise experimental context, toxicity parameters, and publication provenance required to eliminate costly hallucinations. Concurrently, structured ontologies link scientific concepts to patent records and clinical trial results, allowing AI models to operate on deep domain logic rather than superficial text associations.
Grounding enterprise models in peer-reviewed science ensures that discovery tools generate actionable hypotheses anchored in verifiable physical reality.
Sebastian: While AI cannot eliminate paper reading entirely, it fundamentally transforms how researchers navigate vast scientific literature. Instead of replacing human reading, connected data helps teams quickly pinpoint precisely which papers, patents, and trial records deserve their time and focus.
By connecting academic publications, patent filings, grants, and clinical trials into a single semantic graph, researchers can:
Rather than wading through thousands of irrelevant search results or manually combing through PDFs, research divisions use unified knowledge networks to track a scientific concept from early grant funding through to clinical outcomes, ensuring their time is spent analysing the precise evidence that matters.
Sebastian: For pharmaceutical and life science enterprises, some of the most valuable insights lie at the intersection of proprietary laboratory data and external research.
Internal discovery records, assay results, and compound libraries often sit isolated from global scientific literature, international patent registries, and clinical databases. Connecting these disparate ecosystems is essential to avoid duplicating research, uncover hidden target opportunities, and accelerate time-to-market.
Digital Science addresses this challenge through a unified semantic layer . By establishing common ontologies and linked data structures across both internal corporate files and external datasets (like Dimensions ), the semantic layer bridges data silos. Researchers can query internal proprietary discoveries alongside global publications and patent filings seamlessly, revealing critical connections and market opportunities that would otherwise remain hidden in isolated systems.
Sebastian: Staying ahead in R&D requires shifting from reactive literature tracking to proactive landscape monitoring and technology foresight. Modern research divisions combine global publication data from Dimensions with global patent intelligence to continuously map competitive and technological landscapes.
This combined intelligence allows R&D leaders to spot emerging scientific trends, identify fast-growing technology domains, and assess competitor activity long before commercial products reach the market—enabling strategic investments and proactive IP positioning.
Sebastian: Trust in AI systems requires complete transparency—especially in technical fields where legal and scientific accuracy is paramount. Black-box models that generate assertions without showing their work cannot be relied upon for strategic R&D decisions or patent applications.
To build genuine trust across multidisciplinary teams, enterprise discovery platforms need to leverage a semantic layer that enforces full end-to-end auditability. Every insight, graph connection, or automated summary generated by the system is explicitly linked and traced directly back to its original source record—whether that is a peer-reviewed journal article, a clinical trial filing, or a specific patent claims page.
By allowing researchers and patent attorneys to click directly through to the underlying primary evidence, scientific teams gain complete confidence in the integrity and provenance of the system’s outputs.
Digital Science is a key sponsor of this year’s AI & Big Data Expo Europe .
See also: Germany and NVIDIA sign industrial AI declaration in Berlin
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including IoT Tech Expo and the Cyber Security & Cloud Expo . Click here for more information.
AI News is powered by TechForge Media . Explore other upcoming enterprise technology events and webinars here .
