Summary of key findings
1
The first AI wins are here — embedded in scientist workflows and built on trusted data
A handful of tools have broken out of pilot mode. The breakthrough use cases — literature review (76% adoption), protein structure prediction (71%), scientific reporting (66%), and target identification (58%) — succeed due to clean, verifiable data that fits naturally into scientists’ daily work.
2
AI hits a ceiling in complex, regulated science
Adoption drops in areas like generative design, biomarker analysis, and ADME, where data is scattered, incomplete, and hard to validate. Overcoming these gaps is critical as teams look beyond task-level copilots toward systems that coordinate experiments and decisions end-to-end. The biggest areas of planned …
Summary of key findings
1
The first AI wins are here — embedded in scientist workflows and built on trusted data
A handful of tools have broken out of pilot mode. The breakthrough use cases — literature review (76% adoption), protein structure prediction (71%), scientific reporting (66%), and target identification (58%) — succeed due to clean, verifiable data that fits naturally into scientists’ daily work.
2
AI hits a ceiling in complex, regulated science
Adoption drops in areas like generative design, biomarker analysis, and ADME, where data is scattered, incomplete, and hard to validate. Overcoming these gaps is critical as teams look beyond task-level copilots toward systems that coordinate experiments and decisions end-to-end. The biggest areas of planned AI growth — workflow orchestration, manufacturing optimization, multimodal models, and co-scientists — reflect this shift.
3
Scientists have shifted with AI, the infrastructure needs to catch up.
AI has become scientists’ default interface with 89% using copilots or reasoning tools as their first stop to interrogate and synthesize data. As reliance on external data increases, scientists have come to expect open-source tools with flexible access.
4
Builder culture is taking root
The top source of AI talent comes from internal upskilling (67%) and not from tech (21%). Leading organizations are running interdisciplinary sprint groups that test, validate, and fail fast, embracing a “build what differentiates, buy what scales” mindset.
Data sources & methodology
This report draws on a November 2025 survey of ~100 biotechnology and pharmaceutical organizations actively using AI across research and development (R&D).
Importantly, this is not a general industry sentiment study. It is a focused view into the practices and priorities of biotech’s AI leaders and front-runners, organizations that are already deploying AI regularly and shaping how it’s operationalized in modern R&D. Throughout the report, we’ve integrated qualitative insights and emerging best practices from AI and technology experts, biotech industry executives, and early adopters to provide context on where the field is heading.
All respondents are based in the U.S. and Europe, and represent a mix of scientists, technologists, and executives working in or directly supporting one or more of the following functions: discovery research, process and analytical development, bioanalytical science, and animal safety and toxicology. All respondents use AI in their organizations today. The survey was conducted by an independent research firm and expert network to ensure objectivity and industry relevance.
Get the biotech AI guide
At Benchling, we’re helping leading biotech companies implement AI to accelerate their R&D. Here are their top 6 use cases for AI, complete with example prompts that you can try today.
“
We’re seeing AI make its biggest impact in pharma when technology, science, and process design work in unison. At BMS, AI is already supporting nearly every facet of our work. Our integrated approach connects data, AI/ML, wet lab, and clinical expertise into one ecosystem, where insights continually inform decisions, accelerate learning, and help us discover and develop new medicines.
“
We’re at a moment where computation is unquestionably here to stay in drug discovery, and the responsibility now is to balance excitement with rigorous validation. Every company is using AI in some way; what matters is how well they connect those capabilities across the organization. The biggest challenge is talent — people who can navigate both science and ML — but the progress we’re making as an industry is remarkable.
Where to start with AI in biotech
Get 6 practical use cases with example prompts built specifically for biotechnology here.
Not sure where to start with AI?
From structuring PDFs and CSVs, to drafting notebook entries, this guide shares 6 practical use cases for AI with example prompts, built specifically for biotechnology.
“
The models are getting much better at R&D-related tasks, to the point that for many workflows, the bottleneck is in the product layer. Most organizations are still running workflows across dozens of product surfaces, with data fragmented across different locations. The next evolution is a unified interface for AI, connecting continuous model improvements to the experimental data that drives R&D.
Get the biotech AI guide
Read the most common use cases we’ve gathered while talking to hundreds of biotech teams that are using AI. Complete with example prompts you can try today, this guide gives you practical examples to help you get started accelerating your R&D with AI.