In partnership with Thoughtworks we conducted a poll among the 300 delegates at our 2026 Spring Conference to understand the impact AI is having in life sciences R&D. While AI use is becoming widespread across our industry our poll shows AI is not truly embedded across organizations and the value is not reaching into specialized R&D activities. As AI evolves as a capability, collaboration will be essential for closing the gap between AI investment and value. Successful adoption will require getting both the data and people elements right and ensuring we continue to work together alongside regulators to shape the data and standards required to scale AI safely and effectively.
Resource Tag: Harnessing AI To Expedite R&D
Natural language querying of biological databases with large language models
Meaningful querying biological databases today requires specialist knowledge of structured languages such as SQL, SPARQL, or Cypher — making data mining slow, labor-intensive, and inaccessible to many researchers. Reliable natural language querying would change that. It would also be a prerequisite for the next generation of AI co-scientist systems: tools that automate scientific hypothesis generation at scale.
This peer-reviewed paper, produced by the Pistoia Alliance’s Large Language Models in Life Sciences project and published in Drug Discovery Today (May 2026), reports the outcomes of a systematic assessment of current practices in natural language querying with LLMs.
Highlights
- Accurate natural language data mining is a requirement for AI co-scientist systems
- Multi-agent LLM systems combined with deterministic queries offer the best accuracy-flexibility balance
- LLM agents must be Findable and Reusable (FAIR) and require open API standards
- Shared benchmarks for natural language data mining systems are needed across the industry
Authors
Vladimir A. Makarov (Pistoia Alliance), Oleg Stroganov (Rancho Biosciences), Laura I. Furlong (MedBioInformatics Solutions), Brian Evarts (Crown Point Technologies), Loes van den Biggelaar (The Hyve), Alexandros Goulas (AbbVie), Etzard Stolte (Roche), Derek Marren (AstraZeneca), and Lars Greiffenberg (AbbVie).
__________________________________
What did the study test?
The authors evaluated 21 different strategies for translating natural language into structured database queries, using the Open Targets Platform as a real-world target discovery and validation use case. Five LLMs were tested — GPT-4o, Claude 3.5 Sonnet, o1, GPT-4o-mini, and open-mistral-7b — across naïve, template-based, retrieval-augmented, prompt-optimized, and multi-agent approaches.
What works — and what doesn’t?
Naïve prompting failed almost universally on complex scientific questions, even when the full database schema was supplied. Template-based strategies achieved 100% accuracy but are rigid: they cannot be transferred to new data sources without substantial human effort, and they do not scale.
Multi-agent strategies — in which multiple LLM agents challenge one another’s outputs and interact with a human user — achieved 83–98% accuracy on complex questions with the best-performing model (o1).
This combination of accuracy, flexibility, and adaptability across data sources represents the most practical path forward identified in the study. Entity recognition (for example, correctly resolving “ALS” to “Amyotrophic Lateral Sclerosis”) was the single largest remaining source of error.
What does the industry need next?
The paper makes three calls to action for the field:
- Combine multi-agent LLM systems with deterministic API calls for simple, high-confidence retrieval tasks.
- Develop a shared industry benchmark for natural language data mining that is resilient to LLM background-knowledge contamination — a subtle but significant source of evaluation error.
- Establish FAIR-aligned open standards — such as the Model Context Protocol — for the discovery of and engagement with AI agents across commercial and academic systems.
This research was coordinated by the Pistoia Alliance Large Language Models in Life Sciences project. To learn more about the project, visit https://pistoiaalliance.org/project/benchmarks-for-natural-language-data-mining-with-llms/.
Seminar Review: Life Science AI Exchange – March 2026
The Pistoia Alliance Life Science AI Exchange seminar (March 2026) explored how the life sciences industry can better measure, benchmark, and realise the value of AI. Speakers from academia, pharma, and technology highlighted rapid advances in AI capabilities—particularly in scientific reasoning—while emphasising persistent challenges around evaluation, data quality, and overconfidence in model performance.
The session underscored the importance of rigorous benchmarking, combining AI with structured data approaches like knowledge graphs, and adopting practical frameworks to assess real business impact. Industry insights revealed a significant gap between AI ambition and organisational readiness, with skills, governance, and interoperability emerging as key barriers. Overall, the seminar highlighted the need for more disciplined measurement, stronger collaboration, and increased AI literacy to enable reliable and scalable adoption across the sector
Advancing AI & ML In Life Sciences R&D Whitepaper 2025
From Model Innovation to Organizational Readiness
The Next Phase of AI in Life Sciences The findings presented in this report reflect the state of thinking across the life sciences sector during the second half of 2025. In a field where technological cycles increasingly compress into months rather than years, it is important to acknowledge the temporal context. The pace of AI advancement continues to accelerate. Organizational capability evolves more gradually. The widening distance between these two curves is central to the conclusions of this report.
This global survey of 54 senior leaders across pharmaceutical, biotech, CRO, and technology organizations, followed by in-depth interviews with 9 executives, reveals a consistent pattern: internal demand for AI and machine learning is high, perceived strategic importance is high, yet organizational maturity remains comparatively low.
What is the role of the knowledge graph in the age of AI?
AI is reshaping biopharma R&D, yet its effectiveness depends on the structure and integrity of the underlying knowledge layer. Knowledge Graphs are often presented as one of the key approaches — but where do they genuinely deliver advantage, and where do complexity and cost outweigh benefit? This webinar will examine the practical realities of deploying Knowledge Graphs at scale, and consider whether a focused, pre-competitive Pistoia Alliance collaboration could help define shared semantic foundations for the next phase of data-intensive life science research.
Speakers & Facilitator
Facilitator: Brian Martin, Chief AI Product Owner & ACOS Senior Research Fellow, AbbVie
Demand-Side Perspective: Christian Senger, Lead Semantics Technology Expert, Bayer
Engineering Perspective: Ryan Chandler, Knowledge Graph Engineer, AbbVie
Supply-Side Perspective: Tine Geldof, ONTOFORCE
Click here to discover more about our AI Community.
Illustrations of Value Delivered by AI: Benchmarking & Ongoing Value Measurement
How do we move beyond the hype and rigorously demonstrate the value AI delivers in scientific and enterprise settings? This webinar brings together leading voices from academia, pharma, and biotech to tackle one of the most pressing questions facing organisations adopting AI today.
Our expert panel will explore three complementary perspectives:
Benchmarking Chemical Reasoning in LLMs — How do large language models really perform when faced with specialist scientific reasoning? We’ll look at new approaches to structured evaluation that reveal both the capabilities and limitations of AI in chemistry and related domains.
Evaluating Biomedical NLP in Practice — Claims of impressive zero-shot performance don’t always hold up under scrutiny. We’ll examine findings from large-scale evaluations of language models on biomedical text tasks, and explore what the gap between claimed and actual performance means for real-world deployment decisions.
Comparing Value: AI Agents vs Traditional ML — AI agents and traditional data science require very different skillsets, deliver different kinds of value, and demand different ROI benchmarks. We’ll present a practical framework for how organisations can assess and compare these distinct approaches when making investment decisions.
Whether you’re building the case for AI adoption, evaluating tools for deployment, or looking to establish robust measurement practices, this session will equip you with concrete approaches to understanding — and communicating — the genuine value AI brings to life sciences and beyond.
Life Science AI Exchange Launch Meeting
Summary notes of the Life Science AI Exchange roundtable held on February 25, 2026
How to Accelerate Biopharma Research with AI Scientific Agents – Japanese Subtitles
AI tools for supporting scientific research have come a long way over the past couple of years. From limited beginnings, they’re now able to play a key role in accelerating pharmaceutical research. But finding and using the right AI scientific assistant is still a minefield. In this webinar, Rob Brown, Head of Scientific Office at Sapio Sciences, looks at the advancements in AI scientific assistants and agentic AI, how these tools can be integrated into the research process, and what Sapio’s own AI scientific assistant, ELaiN, is capable of.
Speaker
- Rob Brown, Global VP Product and Pre-Sales, Sapio Sciences
Co-Host
- Yuri de Lugt, Global Field Marketing Director, Sapio Sciences
2026 Life Sciences Outlook
Life sciences organizations are entering a new phase where scientific, technical, and regulatory forces are converging. As teams prepare for the next wave of transformation, 2026 is poised to bring meaningful shifts in how discovery, development, and clinical groups design studies, generate insights, and operationalize advanced analytics at scale. Join Domino Data Lab and AstraZeneca for a forward-looking discussion on the emerging trends that will shape the year ahead and what leaders can do now to position their teams for impact across both AI and modern data science practices.
Why watch
- Hear AstraZeneca’s perspective on the trends that will matter most for data science and AI across R&D, clinical development, and digital transformation in 2026
- Get practical guidance on how data science and IT teams can prepare for rising expectations around model governance, analytical rigor, and operational readiness
- Learn how peers are planning for 2026 and what capabilities in data management, scientific computing, and AI-enabled workflows will differentiate organizations entering the next stage of life sciences innovation
Speakers
- Jamie O’Keefe, Head, Development and Regulatory Data, Digital & Technolody, Takeda
- Tiffany Fabianac, Director – FAIRR Analytics, Oncology Data Science & AI, AstraZeneca
- Chris McSpiritt, VP of Life Sciences Strategy, Domino Data Lab
Executive Summary: Pistoia Alliance Agentic AI Workshop
The Pistoia Alliance’s Agentic AI Workshop held on November 11, 2025 brought together leaders from pharma, biotech, automation, and frontier AI to showcase how multi-agent systems are already accelerating R&D, clinical workflows, and quality processes across the life sciences. Speakers from Genentech, Ginkgo Bioworks, Owkin, Ketryx, Roche, Autopoiesis, and Google DeepMind demonstrated early production deployments, revealing major gains in speed, consistency, and scientific insight—while also emphasizing shared challenges around governance, data quality, hallucinations, traceability, and wet-lab integration. Across presentations and panel discussions, a clear consensus emerged: agentic AI is rapidly moving from experimentation to enterprise impact, and the industry now needs common standards, robust oversight frameworks, and collaborative ecosystems to safely scale these systems across regulated environments.
AI Session Summary
As part of the Pistoia Alliance’s 2025 Fall Conference, this session brought together experts from Pfizer, Elsevier, Genentech, Rancho Bioscience, AstraZeneca, Domino Data, and the Pistoia Alliance to examine the accelerating impact of AI and agentic systems in life sciences. Speakers highlighted the limitations of standalone LLMs, the need for agents to access live knowledge, and the importance of ontologies for context and explainability, alongside the practical challenges of deploying multi-agent systems within regulated environments. Core themes included trustworthy data, governance, workflow integration, and the growing shift toward AI as a co-researcher. The panel stressed that safe, effective adoption will require pre-competitive collaboration, transparent standards, and deeper engagement with regulators to ensure AI systems are traceable, reliable, and aligned with scientific and patient outcomes.
AI Isn’t the Product, the User Experience Is (Japanese Subtitles)
Brought to you by the AI workstream of the UXLS community
In pharma R&D, UX design is the difference between an AI breakthrough that changes how scientists work and a promising model that never leaves the lab.
Exploring how the future of AI in pharma R&D depends not just on the technology itself, but on the experience of the people using it, this webinar brings together voices from the pharmaceutical industry, software creators, and UX professionals. The discussion will focus on humans at the core of AI adoption and how considered design transforms AI from a promising tool into a breakthrough that scientists and businesses can truly embrace.
こちらは、UXLSコミュニティの AI ワークストリームがお届けします。
製薬 R&D において、UX デザインは「科学者の働き方を変える AI のブレークスルー」と「研究室から出られない有望なモデル」を分ける重要な要素です。
本ウェビナーでは、AI の未来がテクノロジーそのものだけでなく“それを使う人の体験”に左右されるという視点から、製薬業界、ソフトウェア開発者、UX の専門家が集まり議論します。AI 導入の中心にいる“人”に焦点を当て、熟慮されたデザインが AI を単なる有望なツールから、科学者や企業が真に受け入れられるブレークスルーへと変える方法を探ります。