Artificial intelligence is still struggling to crack the most complex problem in pharma R&D: reliably designing new molecules that become breakthrough medicines. But drugmakers say AI is already paying off in the “unsexy” middle of development. This is where time and money are often lost to paperwork, vendor hand-offs, and slow trial start-up.
At the JPMorgan Healthcare Conference in January 2026, executives from large pharma companies and smaller biotechs described how AI tools are accelerating trial planning, participant recruitment, and regulatory submission preparation. Several companies said these gains are shaving weeks off labour-heavy processes. However, the impact on overall drug approval timelines will take longer to prove.
AI in Drug Development is Shrinking Regulatory Work
One of the clearest early wins is regulatory documentation. Drugmakers must assemble and maintain thousands of pages covering clinical, safety and manufacturing data. These packages must remain consistent across regions and formats. AstraZeneca has noted that this work often results in heavy reliance on external contractors. This pushes up costs and adds risk when information is duplicated or inconsistent.
AI is being used to draft, cross-check and standardise documents, and to convert reports into regulator-ready templates. Germany’s radiopharmaceuticals firm ITM, for example, has described using AI to reshape lengthy trial reports into US FDA-style formats. This potentially saves weeks of staff effort once deployed.
AI in Drug Development is Speeding Clinical Trial Set-Up
Trial site selection and enrolment are also being compressed. Novartis said it used AI in 2023 to launch a 14,000-person cardiovascular outcomes study for its cholesterol medicine Leqvio. A process that typically took four to six weeks became a single two-hour meeting. As a result, the company could pick higher-performing sites and close enrollment near the target. Novartis has said these time savings can add up to months across a complete development programme.
British drugmaker GSK said its mix of digital and AI tools is designed to accelerate clinical trials by 15%. The company has linked the approach to about £8 million in savings in late-stage studies of its twice-yearly asthma drug Exdensur. This drug won US approval in December 2025.
The “Messy Middle” is Where Investors Want Proof
Venture and industry leaders are increasingly focused on what one investor called drug development’s “messy middle”. Andreessen Horowitz’s Jorge Conde has described trial enrolment as a “leaky funnel”, where participants drop out during outreach, screening and scheduling. He has backed the startup Alleviate Health, which applies AI to patient education, engagement, and logistics.
Analysts say admin-focused AI adoption is already common. TD Cowen’s Brendan Smith has noted that tools such as Microsoft Copilot are being used for routine tasks. Still, investors may need one to three years before the speed gains show up clearly in measurable programme metrics.
Agentic AI and Partnerships are Expanding Fast
Momentum is growing around “agentic AI”, which can execute workflows with limited human input. McKinsey projects that agentic AI could increase clinical development productivity by 35% to 45% over the next five years.
Several companies are also partnering aggressively. Eli Lilly has highlighted its work with Nvidia as part of broader bets that AI can improve development success rates. Genmab, meanwhile, has announced a partnership with Anthropic to deploy Claude-powered agentic AI to support clinical development. This includes automating post-trial analysis and converting outputs into figures and clinical study reports.
For now, the message from industry leaders is pragmatic. AI may not have yet delivered the first universally recognised “AI drug”. But it is already reducing friction in trials and regulatory operations, and freeing skilled teams to focus on higher-value scientific work.