SAN FRANCISCO, January 26, 2026, 05:15 (PST)
- Drugmakers say AI is cutting time spent on trial logistics and regulator-facing paperwork
- Novartis said AI shortened trial site selection for a 14,000-person Leqvio study to a two-hour meeting
- TD Cowen’s Brendan Smith said it could take 1–3 years before investors can gauge the impact
Drugmakers are turning to artificial intelligence to speed up parts of clinical trials and regulatory submissions, cutting weeks from tasks that usually bog down development, executives and analysts said. The technology has not yet delivered the big breakthrough — finding new molecules that become major medicines — but it is already biting into the “messy middle” of drugmaking. (Reuters)
That matters because the clock is still brutal. Companies say it can take about a decade and roughly $2 billion to bring a drug to market, and the slow grind of trial setup and documentation can stretch timelines even when results look promising.
A wave of AI deals has followed, including Eli Lilly’s partnership with chipmaker Nvidia, as drugmakers test tools meant to lift success rates and shrink timelines. McKinsey last year said “agentic AI” — systems that can complete tasks with little human intervention — could raise clinical development productivity by about 35% to 45% over the next five years.
Teva CEO Richard Francis framed the early wins as boring, by design. “Everything else” around drug development needs to get “as efficient and as small as possible,” he said, pointing to digitization and process work that “makes a difference.”
Paperwork is a major target. Companies including AstraZeneca have described tracking thousands of pages for regulators, spanning clinical, safety and manufacturing records, and keeping them consistent across regions.
AstraZeneca Chief Financial Officer Aradhana Sarin said teams often have to compile, cross-check and harmonize those documents across geographies, sometimes leaning on outside contractors to do it.
Some of the money is chasing what investors call bottlenecks, not breakthroughs. Andreessen Horowitz general partner Jorge Conde said he is backing fixes to drug development’s “messy middle,” including a $4.3 million investment in startup Alleviate Health.
Conde called trial enrollment a “leaky funnel,” with would-be participants dropping out along the way, and said Alleviate is using AI to support outreach, education, screening and scheduling.
TD Cowen analyst Brendan Smith said large language models — chat-style AI systems trained on text, such as Microsoft’s Copilot — are becoming common for administrative tasks in pharma. But he said it may take another one to three years before investors can measure how much the tools are accelerating drug development, and quantifying savings will vary by how companies deploy them.
Novartis pointed to a concrete example. Chief Medical Officer Shreeram Aradhye said the company used AI in 2023 as it launched a 14,000-person late-stage outcomes trial for cholesterol drug Leqvio — late-stage meaning one of the final large tests in patients before a filing for approval.
Aradhye said AI cut the typical four-to-six-week process of picking trial sites — selecting hospitals and clinics to run the study — into a two-hour meeting, and helped Novartis end enrollment with only 13 patients above its target. “AI becomes augmenting intelligence, not artificial intelligence,” he said.
GSK said it is using a mix of digital and AI tools to reduce manual data collection and aggregation and to improve enrollment, with an aim to speed up all clinical trials by 15%. A spokesperson said that saved about 8 million pounds ($10.87 million) in costs for late-stage studies of asthma drug Exdensur last year, and the drug won U.S. approval last month. (The Economic Times)
Genmab said it plans to deploy Anthropic’s Claude chatbot-powered agentic AI to support clinical development, including automating post-trial work such as turning data into graphs, tables and clinical study reports, according to its head of AI Hisham Hamadeh. German radiopharmaceuticals firm ITM said it has worked out how to use AI to convert long trial reports into U.S. Food and Drug Administration template formats, potentially saving weeks, but has not yet deployed it.
Amgen research chief Jay Bradner said AI is already helping drug development and regulatory document preparation, even as the industry waits for proof it can deliver new medicines end-to-end. “What everybody’s waiting for is the AI drug,” he said. “I actually think those molecules are in pipelines right now.” (Investing)
But the case for AI as a true time machine is still forming. Smith said it could take years before investors can see whether shorter site selection, faster enrollment and cleaner submissions add up across a program, and some companies say tools they have built are not yet in use.
