Reimagining R&D with AI
Bringing a new drug to market is one of the most high-stakes processes in business and science, requiring an average of ten years and more than a billion dollars. Meanwhile, the number of medicines in the R&D pipeline has doubled over the past decade, intensifying the pressure to innovate.
The biopharma industry has responded by rapidly exploring the potential of generative AI to modernize critical assets and workflows throughout the R&D lifecycle. This week, McKinsey is proud to share a milestone in that journey: collaborating with Merck to streamline a manual, time-consuming step in drug development—the drafting of clinical study reports (CSRs). Together, we have developed a platform that allows teams to reduce the time it takes to complete CSRs from two to three weeks to three to four days, while improving accuracy.
“These complex, critical-path documents are submitted in the final phase of the drug approval process,” explains Matthew Wilson, a senior partner at McKinsey. “By using the power of gen AI to automate the first draft, we’re hopeful that we’ll start to see medicines in the hands of patients faster.”
An AI-powered collaboration: technology, transformation, and talent
Over the past 18 months, McKinsey and Merck have co-developed an innovative authoring platform for CSRs that draws on the deep AI and data science expertise of QuantumBlack, McKinsey's AI arm. But the effort doesn’t stop at technology. Together, we reimagined Merck’s operational processes, redesigned roles, and upskilled medical writers and tech professionals. The results? The team has:
- reduced the time to create a fully human-reviewed CSR first draft from an average of 180 hours to 80 hours
- increased the quality of CSR drafts–as measured by reducing the number of errors by 50%–in categories such as data, messaging, citations, terminology and typography.
QuantumBlack, AI by McKinsey
Learn how we're helping capture the opportunity of generative AI
CSRs are traditionally labor-intensive, requiring teams of medical writers to sift through thousands of pages of clinical data over a period of months. Writers must resolve conflicting data, verify results, and construct narratives using specific regulatory language.
Despite widespread interest in applying gen AI to streamline this work, companies have faced challenges building platforms that generate full CSRs of regulatory quality. Earlier this year, our joint team of more than 80 data science, AI, life science, and medical experts and technologists, working across three continents, produced and submitted the first live CSRs using the new proprietary platform.
The solution integrates advanced table preprocessing with gen AI authoring. It handles table mapping, data extraction, styling, and validation, generating a high-quality first draft in just five minutes. It covers 14 required CSR sections for Merck's highest volume Therapeutic areas. “Our team worked with Merck in agile sprints bringing the latest of gen AI techniques and a suite of pre-built accelerators,” shared Chaitanya Adabala Viswa, a partner with QuantumBlack. “Our deep collaboration allowed us to innovate seamlessly, delivering impact with trust.”
From wrangling data to generating insights
Crucially, the platform is designed to operate with a medical writer in the loop, marrying the speed and accuracy of AI with the clinical judgment and expertise of experienced professionals to create a low-risk, high-reward workflow.
“Our innovative work in Information Technology is focused on getting our products to patients faster,” said Matt Studney, senior vice president of Merck Research Laboratories Information Technology. “This work puts AI in the hands of our medical writers and across our enterprise, creating new efficiencies for our skilled R&D teams, who are integral to our innovation and success.”
I co-wrote a CSR that took months of preparatory work and then weeks of authoring. To see a CSR pop up on your screen in eight minutes is amazing. We can work so much smarter with this.
For medical writers, the shift is transformative. Freed from hours of wrangling and formatting manual data, they can focus on developing insights through analysis.
As one principal scientist observed, “I co-wrote a CSR that took months of preparatory work and then weeks of authoring. To see a CSR pop up on your screen in eight minutes is amazing. We can work so much smarter with this.”
Scaling the future of clinical documentation
Merck is in the process of scaling the platform across its late phase pipeline by the end of the year. The teams are building a library of prompts and reusable components, shifting mindsets toward fit-for-purpose drafting to address the end-to-end submission journey.
“Our LifeSciences.AI platform includes a broad set of industry AI capabilities, including authoring multiple types of documents,” says Jeffrey Lewis, a McKinsey senior partner. “We are excited about our collaborations that use the power of AI–predictive, generative, and agentic–to push the boundaries of what is possible across the entire R&D cycle to bring new medicines to patients faster.”