AI is shifting from a data tool to a core pillar of pharmaceutical R&D, according to Thorsten Rall, Global Life Sciences Leader at Capgemini. AI platforms and closed-loop automated laboratories are compressing discovery timelines, improving clinical trial design, and delivering higher-quality drug candidates faster than traditional methods.
Quality over speed in drug discovery
Rall sees the biggest transformation in target discovery, hit and lead prioritisation, and in silico experimentation that cuts down on laboratory and animal testing. He said, "For me, the biggest opportunities are in target discovery, hit and lead prioritisation, and in silico experimentation that can reduce reliance on some forms of laboratory and animal testing."
A recent Capgemini Research Institute report found that 63% of biopharma executives believe most new molecular entities will originate from AI-driven platforms within a decade. While AI can accelerate parts of the discovery process, Rall argues the more significant impact is on candidate quality. Historically, drug discovery screened enormous numbers of molecules and tolerated high failure rates. AI now lets teams identify more promising targets earlier and design molecules far more efficiently.
Insilico Medicine's average timeline to nominate a preclinical candidate is nine to 12 months, compared with the industry average of four to six years. The company also demonstrated how AI can pinpoint previously overlooked disease drivers. For idiopathic pulmonary fibrosis, Insilico applied large-scale biological data analysis to identify TRAF2- and NCK-interacting kinase (TNIK) as a novel target, reducing the path from target identification to clinical candidate to under three years. Early clinical data showed encouraging safety and efficacy signals, and Rall sees such examples as evidence that AI is lifting the quality of targets and candidates entering development.
Closed-loop labs connect wet experiments with in silico models
Most organisations still treat AI as separate from the laboratory, but Rall points to the strongest results when AI integrates directly into physical experimentation. Capgemini's deep tech group Cambridge Consultants combines wet lab work with in silico environments, allowing researchers to design, run, and refine experiments with far fewer data points.
Rall said the approach becomes especially powerful in closed-loop systems where wet labs, computational models, and AI learn from each other continuously. Experimental results feed into models that then suggest the next round of tests. At Cambridge Consultants, the team achieved the brightest expression variant of a Green Fluorescent Protein using only 46 data points, versus an estimated 80,000 needed through directed evolution. The experiments then generated new data that improved the models further.
AI makes clinical trials faster and more adaptive
The Capgemini research indicates that 60% of R&D leaders expect AI to substantially improve clinical trial efficiency. Immediate gains are showing up in patient recruitment, site selection, and operational management. AI helps identify eligible participants more quickly, refines trial protocols, and compresses clinical statistical analysis from weeks to days.
Longer-term, Rall sees synthetic and simulated data reshaping how clinical evidence is generated. Synthetic control arms could reduce reliance on placebo groups for severe or rare diseases, a move that carries both scientific and ethical weight. One study recreated a control arm from past trial data and real-world registries, matching the survival outcomes of the original randomised trial. Before such methods enter routine use, however, regulators will need to build confidence in the underlying models and evidence.
AI agents become scientific thought partners
Rall envisions AI agents acting as companions for scientists, helping them challenge assumptions and discard weaker hypotheses earlier. "I see AI agents as companions for scientists," he said. Already 38% of organisations are piloting AI agents in R&D, according to the Capgemini survey.
To reach their full potential, these agents need specialised models trained on scientific data, grounded in biological and molecular understanding, and capable of explaining their reasoning. Strong data foundations, governance, and trust in AI outputs are just as important as the technology itself. Researchers must have confidence that a target recommendation is based on complete, high-quality data and can be validated.
Why this matters for science and research professionals
Rall expects AI to transform pharmaceutical R&D more than any technology in decades. In the near term, the industry will see faster, more efficient research processes. Over time, AI platforms will pursue biological targets once considered too complex, identify new patient populations for existing therapies, and deepen the understanding of disease mechanisms. "I believe AI has the potential to transform pharmaceutical R&D more than any technology we have seen in decades," Rall said.
For scientists and researchers, the shift means working with AI not as a separate tool but as an embedded part of experimental design, target identification, and clinical strategy. The ability to design closed-loop experiments and interpret model-driven hypotheses is becoming a core skill. As disease-agnostic AI platforms emerge, professionals who can collaborate effectively with these systems will shape the next generation of medicines.
Your membership also unlocks:
