GraphTalk Pharma & Life Sciences returned to Munich on 2 July 2026 as a hybrid event, bringing together pharma, biotech, and life sciences leaders to explore how knowledge graphs and GenAI move organizations from AI pilots to production.
Structured around the pharma product lifecycle, the day covered R&D, clinical, and supply chain use cases, with speakers from Bayer, KWS, BASF, Sandoz, Boehringer Ingelheim, Insel Gruppe, SwissDRG, and Neo4j sharing real-world applications of context graphs, GraphRAG, and agentic AI. Attendees left with concrete examples of how connected, explainable data is helping pharma and life sciences organizations turn fragmented information into grounded, trustworthy AI-driven decisions.
The morning's R&D track showed how Bayer and KWS Saat are using knowledge graphs to accelerate early-stage discovery, from gene-disease hypothesis generation to explainable candidate gene identification in plant breeding, while BASF shared how its "Plant Twin" graph consolidates fragmented biological data for non-model species. The clinical track shifted focus to healthcare data quality, with Insel Gruppe demonstrating how free-text Swiss medication records can be mapped into a structured, queryable graph, and Sandoz showing how patent data modeled as a graph exposes patterns of delayed generic and biosimilar competition.
Two streaming sessions rounded out the clinical block, covering a newcomer's experience building a Swiss drug data graph and a temporal knowledge graph approach to clinical decision support.
In the afternoon, the supply chain track brought a sharper commercial edge: Boehringer Ingelheim presented an agentic layer that lets non-technical users query a unified pharma supply chain graph in plain language, and Amalgo quantified the financial upside of cutting QA investigation times using context graph simulations. The day closed with a technical deep dive from Neo4j's own team on preparing context graphs for agentic and visual intelligence, tying the day's use cases back to a common architectural foundation. Across every track, the throughline was consistent: knowledge graphs are what let GenAI move from impressive demos to auditable, production-grade systems that pharma and life sciences organizations can actually trust.
Vladislav Kim presented two knowledge-graph-driven methods for early drug target discovery: graph reasoning to infer novel gene-disease associations, and GraphRAG to anchor retrieved facts to verifiable graph relations. Together, the approaches reduce hallucination and strengthen both predictive power and grounding for AI-supported target hypothesis generation.
KWS Saat:
From graph to gene: integrating Knowledge Graphs and AI for explainable candidate gene discovery
Bjoern Oest Hansen described a framework combining knowledge graphs, link prediction, and generative AI to identify candidate genes for complex agricultural traits. By integrating genes, traits, and biological evidence into a single graph, the approach produces biologically meaningful, explainable hypotheses rather than opaque predictions.
Brent Murphy walked through BASF's "Plant Twin" graph, which consolidates omics data, orthology, pathways, and literature to fill knowledge gaps for non-model plant species. The project, built in collaboration with Neo4j Professional Services, moved from concept to a working proof of value in roughly six months.
Insel Spital Bern:
From free text to Knowledge Graphs: visualizing Swiss medication and procedure mapping in Neo4j
The team showed how they harmonized multilingual, free-text Swiss electronic health record data on medications and procedures into standard vocabularies using Neo4j. The resulting graph supports interactive visualization, gap analysis, and quality assessment for clinical interoperability, using antibiotic resistance tracking as a driving use case.
Dr. Peeyush Sahu explained how patent data — spanning primary protection, secondary protection, and regulatory exclusivity — can extend originator monopolies well beyond a product's core patent life. He showed how modeling European Patent Office data as a graph makes these "delayed competition" patterns visible and analyzable at scale.
Boehringer Ingelheim:
Supply Chain Insights, Without the Training Manual: Agents on a Pharma Context Graph
Mambwe Mumba described how Boehringer Ingelheim's Supply Chain Insights (SCIX) project unified previously siloed supply chain, quality, regulatory, and finance data into a single context graph. He outlined the next step — an agentic, natural-language layer — that lets users trace products end-to-end, assess network risk, and run "what if" scenarios without needing to learn the underlying query interface.
Jonathan W. Lowe quantified the financial impact of cutting biopharma QA investigation times from a 30-day industry standard down to 10 days, citing double-digit swings in gross profit, penalties, and inventory costs. He used knowledge graph simulations to model bullwhip effects across the supply chain and show where faster investigations create the most value.
Dr. Alexander Jarasch opened the day by framing why AI projects stall in pilots: fragmented data and context-free models that hallucinate rather than reason. He showed how GraphRAG grounds AI answers in verified, traceable graph relationships, giving pharma organizations the regulatory-grade explainability needed for GxP, EMA, and FDA compliance.
Jean-Marc Guerin & Niels De Jong closed the day with a hands-on look at how Neo4j combines knowledge graphs, agentic AI, and graph visualization to ground AI in trusted enterprise data. They demonstrated how GraphRAG and agentic workflows, paired with Neo4j Studio, let users explore and validate AI-generated insights interactively.
» Life Sciences Developer Hub: Explore healthcare and life sciences use cases, demos, technical guides, and hands-on resources.
» IDC Report – The Business Value of the Neo4j Graph Intelligence Platform: Discover how organizations achieved an average 230% ROI while accelerating AI, analytics, and business outcomes with connected data.
» Ready to Try Neo4j? Get Started for Free
AuraDB Free is the easiest and fastest way to build graph applications in the cloud. It’s great for small development projects, learning, experimentation, and prototyping.
» Want to Master the Foundations of Neo4j? We Offer Free, Hands-on Online Training
GraphAcademy offers self-paced courses ranging from beginner-level basics to advanced, deeper dives that can help you become a graph expert!
