Executives from major pharmaceutical companies and a number of smaller biotechnology firms told attendees at the recent JP Morgan Healthcare Conference that artificial intelligence is delivering practical time savings on routine but resource-intensive parts of drug development - even if it has not yet fulfilled the promise of identifying novel molecules that become major medical advances.
Speakers described AI-assisted improvements in finding trial participants and clinical sites, drafting and reconciling regulatory submissions, and automating administrative tasks. Seven large drugmakers and six smaller biotech companies reported uses that they say can shave weeks off processes that traditionally absorb extensive staff time and external contractor support.
Pharmaceutical development remains costly and lengthy. Companies say it can take as long as a decade and about $2 billion to move a new drug from concept to market. Some firms, including Eli Lilly - which has entered into a partnership with chipmaker Nvidia - are betting AI can also lift the success rate for candidate drugs. But those expectations for early discovery have not yet produced widely acknowledged, transformative results.
Still, a number of firms emphasized the more immediate benefits of AI-driven digitization and workflow modernization. Teva Pharmaceutical Industries said it is applying AI across multiple functions to concentrate resources on the core objective of bringing therapeutics to patients, with chief executive Richard Francis noting that the surrounding processes need to be "as efficient and as small as possible." He added that what he described as the unglamorous aspects of modernization - process improvement and digitization - are where AI can make a meaningful operational difference.
Executives from AstraZeneca, Roche and Pfizer, along with representatives of smaller companies such as Spyre and Nuvalent, highlighted the burden of managing voluminous regulatory materials. They said clinical, safety and manufacturing records can run to thousands of pages that must be compiled, cross-checked and kept consistent across jurisdictions. According to AstraZeneca chief financial officer Aradhana Sarin, that frequently requires expensive outside contractors to ensure accuracy and alignment across geographies.
Venture investors are targeting what one investor calls drug development's "messy middle." Jorge Conde, a general partner at Andreessen Horowitz, said his firm is backing startups focused on reducing attrition in trial enrollment and improving outreach using AI. He cited a $4.3 million investment into startup Alleviate Health, which aims to address what he described as a "leaky funnel" where prospective participants drop out during outreach, screening and scheduling. Conde said Alleviate applies AI to patient outreach, education, screening and scheduling to reduce that attrition.
Analysts note that many of the administrative AI applications are already becoming commonplace. TD Cowen analyst Brendan Smith said tools such as large language models, including Microsoft Copilot, are being used across the industry for administrative duties. Smith cautioned that investors may need another one to three years before they can assess how AI has measurably accelerated drug development timelines, and that quantifying savings depends heavily on how and where tools are deployed.
Consulting firm McKinsey has published a forecast that more autonomous, agentic AI could lift clinical development productivity by about 35% to 45% over the next five years. Whether that level of productivity gain materializes will depend on adoption and integration of the technologies into existing development processes.
Novartis offered a concrete example of time savings with AI. The Swiss company used AI during preparations for a 14,000-person, late-stage cardiovascular outcomes trial for its cholesterol drug Leqvio. Chief medical officer Shreeram Aradhye said AI reduced site selection from a typical four to six weeks to a two-hour meeting by helping identify higher-performing sites. Novartis reported that it closed patient enrollment with only 13 individuals above its planned target. Aradhye described the technology as augmenting intelligence rather than replacing it, saying, "AI becomes augmenting intelligence, not artificial intelligence." A Novartis spokesperson added that time savings from AI can accumulate to months over the life of a drug-development program.
Other companies outlined targeted deployments and projected savings. GSK said it is using a combination of digital and AI tools to reduce manual data collection, aggregation and trial enrollment times, aiming to accelerate all clinical trials by roughly 15% through these measures. GSK reported that such efforts helped save about 8 million pounds - roughly $10.87 million - in late-stage studies of its asthma drug Exdensur last year. That drug received U.S. approval last month, according to company remarks.
Danish antibody developer Genmab said it intends to deploy Anthropic's Claude chatbot-based agentic AI to support clinical development tasks. Genmab's head of AI, Hisham Hamadeh, said the company plans to automate post-trial activities including the analysis of data and its conversion into graphs, tables and clinical study reports.
German radiopharmaceutical firm ITM told conference attendees that it has developed an approach to use AI to convert lengthy trial reports into U.S. Food and Drug Administration template formats. The company said that approach could save several staff weeks of work but has not yet put the approach into production.
Amgen's research chief Jay Bradner summarized the current state succinctly, observing that AI is providing gains across multiple parts of development and regulatory document preparation. "What everybody's waiting for is the AI drug. When do I get the AI drug?" he said, adding his view that molecules enabled by AI are likely already progressing through pipelines.
Exchange rate note: $1 = 0.7358 pounds.