09/24/2026  |  Pfizer

https://www.pfizer.com/news/articles/how_ai_is_helping_match_patients_to_clinical_trials_faster 

Long before reaching pharmacy shelves, every new medicine must complete one or more clinical trials. But before researchers can determine whether a potential new treatment is effective, they have to find enough eligible volunteers willing and able to participate in a trial. This process represents one of the biggest challenges in drug development.

Patient enrollment is the cornerstone of every clinical trial, but it’s usually the longest phase of the process. Recruiting enough eligible participants can take anywhere from a few months to several years, which can delay the research needed to bring investigational medicines to patients.

As the use of artificial intelligence becomes more common in healthcare, it’s starting to reshape the clinical trial recruitment and enrollment process. AI doesn’t replace researchers or clinicians — rather, it helps them work more efficiently by identifying potential trial participants, improving study design, and uncovering important insights from enormous amounts of data.

“The opportunity here is really to streamline how we work,” says Melinda Rottas, Vice President, Head of Optimization, Analytics and Recruitment Solutions at Pfizer. “Clinical trials are incredibly complex. You have a lot of data, a lot of systems, and a lot of information coming in from different sources. AI can help [our human researchers] compile that information, surface important insights, and ultimately improve both speed and quality.” 

Why enrollment is so challenging
On the surface, it may seem like recruiting participants for a clinical trial should be easy: Find people with a particular condition and connect them with the study that’s right for them. But the reality is much more complicated. Participants in each and every study must meet specific eligibility criteria to help researchers answer questions about the potential new medicine. The safety of each participant is also essential.

“Patient enrollment sits at the intersection of a number of different factors," Rottas explains. “The science itself, how healthcare is delivered, and human behavior — how people feel about research, how they feel about physicians, and the trust they have in the healthcare system.”

For those who may be interested in a clinical trial, being potentially eligible medically is only one step. Prospective participants must first know the trial even exists. Then, they must have access to a participating research site, meet specific eligibility (inclusion and exclusion) criteria, and be able to commit to the required visits and procedures.

The recruitment timeline varies significantly depending on the study. Recruiting participants for seasonal vaccine trials may only take a few months. On the other hand, studies in oncology, rare diseases, or pediatric populations can take three years or more simply to enroll enough volunteers. For patients waiting for new potential treatment options, those delays can be meaningful.

“We run clinical trials to create meaningful results that ultimately lead to regulatory submissions and, hopefully, life-changing treatments for patients,” Rottas says. “Any delay pushes that process and potential progress further back. Many of the conditions we study are serious and can significantly affect a person’s health or quality of life, so every delay impacts our ability to deliver breakthroughs that change patients’ lives.” 

AI helps at every stage
While patient recruitment is one of AI’s most promising applications, Pfizer is also exploring how AI can support nearly every stage of clinical development. The work begins before a study even launches.

Researchers are currently evaluating the use of AI in streamlining protocol development. For example, AI supports our research teams by drafting portions of study documents and providing feedback about whether trial designs might place unnecessary burdens on participants. AI is also being leveraged to help research teams make better-informed protocol decisions using historical research and operational data.

When the protocol is complete, AI has the potential to identify possible sites for the study itself. Sites are typically selected after evaluating multiple factors, including disease prevalence, historical enrollment patterns, healthcare infrastructure, and standard-of-care practices. AI also has the potential to optimize the complex process of activating research sites, coordinating regulatory approvals, training investigators, and ensuring the study medication reaches participating study locations more efficiently.

But perhaps one of AI’s greatest opportunities lies in allowing researchers to identify potential participants faster.1 “What typically happens today is that a person at a clinical trial site manually reviews charts and makes phone calls to identify potentially eligible trial participants,” Rottas says. “Can we accelerate some of that with an AI approach? That’s exactly what we’re working on with research sites today.”

Rather than replace human decision-making, AI works with researchers so they can spend less time searching for participants and more time connecting potentially eligible patients with clinical trial opportunities.

AI doesn’t replace humans
As researchers use AI more, one misconception continues to surface: that algorithms will replace physicians and researchers. According to Rottas, nothing could be further from the truth.

“We talk a lot about having a ‘human in the loop,’ or even ‘human at the helm,’” she says. “AI can be a great thought partner. It can help with research, automate repetitive tasks, generate documents, and surface insights. But there is always a person consuming that information and making the decisions.”

AI allows researchers to focus more of their time on high-value work without replacing their expertise. That idea also extends directly to patient recruitment. Even if AI identifies someone who appears eligible for a clinical trial, the technology doesn’t decide whether they participate.1

“That individual still needs to sign informed consent, go through a robust screening process, and the principal investigator or study doctor will ultimately determine whether that patient is eligible,” Rottas explains. “Maintaining that human in the loop is really paramount.”

Improving access through smarter recruitment 
AI could also help expand access to clinical trials by identifying opportunities that traditional recruitment approaches might miss. In the past, researchers often selected trial sites based largely on previous enrollment performance. AI-enabled patient matching provides a different approach by identifying where eligible patients actually receive care.

“AI-enabled patient matching lets us get very specific about where there might be good opportunities for patients. 

Additionally, researchers can make use of AI to better understand where specific patient populations receive care. It can identify barriers to participation and tailor outreach strategies to different communities. Perhaps most importantly, it has the potential to make clinical trial opportunities more visible and accessible for people who have historically been underrepresented in research.

But technology alone can’t build trust. “Anyone who’s considering a clinical trial is making a big decision,” says Rottas. “AI can help us create more awareness, identify eligible patients, and streamline the process. But now more than ever, we have to ensure people can trust the clinicians, researchers, and study teams responsible for the research they’re considering or participating in.”

That trust is built through community engagement, partnerships with patient advocacy organizations, listening to patient feedback, and ensuring research reflects the needs of the communities it serves.

Looking ahead
Rottas says Pfizer has an effort underway to ensure its data is AI-ready.

“In clinical research, we’ve always relied on rigorous standards for data quality, representativeness, and validation,” she says. “AI doesn't change those fundamentals, it raises the bar. We need to understand where the data comes from, how representative it is, where biases may exist, and how these collective efforts may affect outcomes across different patient populations. Just as importantly, we need strong governance to ensure AI is used responsibly, transparently, and in ways that build trust.”

For more than 175 years, Pfizer has generated scientific and clinical insights across a wide range of diseases and treatments. That expertise, combined with robust data governance and oversight, helps inform how Pfizer responsibly develops and deploys AI solutions. This also ensures there’s enough privacy protections and compliance with evolving regulatory guidance.

Looking to the future, Rottas envisions AI connecting every aspect of clinical development — from protocol design and regulatory strategy to manufacturing, site selection, and patient recruitment — into a more integrated decision-making process. But she does offer one important note of caution.

“I'm excited by the potential of AI, but the foundations of great clinical research remain the same," she says. "We need to meet patients where they are, build trust, expand access to clinical trials, and ensure the people participating in research reflect the populations our medicines and vaccines are intended to treat. AI can help us scale and accelerate those efforts, but it works best when paired with a deliberate focus on patients and communities.”

Ultimately, AI doesn’t change the purpose of clinical trials. Instead, it strengthens the path forward. By helping researchers identify opportunities faster, reducing administrative burdens, and improving patient-centered trial design, AI’s potential could accelerate research while keeping people — not technology — at the heart of every decision.