Artificial intelligence (AI) is increasingly being framed as a transformational tool that can address inefficiencies across therapeutic research and development (R&D). However, upgrading its use beyond heterogeneous pilots while being mindful of potential patient concerns remains a key challenge.
With a wealth of AI-based solutions entering the market, industry experts at the European Clinical Trial Outsourcing 2026 conference shared insights on how these technologies are being incorporated into existing workflows, from accelerating drug discovery to optimizing daily operations.
Discover effective B2B marketing
We combine business intelligence and editorial excellence to reach passionate professionals across 36 major media platforms.
learn more
Currently, AI-integrated workflows are centered around a human-involved approach, with humans reviewing each step of the process, said Piotr Mashlak, senior director and head of emerging technologies at AstraZeneca. The goal is to move toward a human-on-the-loop strategy, where humans oversee workflows rather than having humans evaluate all output, he added. The reliability of AI-based tools will be central to the evolution to more automated processes, Mashlak says.
“Real-world” applications of AI in clinical trials include patient screening, baseline intelligence, data mining, molecular matching, and document intelligence, Mashraq said. Investable strategies for future benefits include agent orchestration and the use of digital twins as a comprehensive control arm, he added. However, other applications, such as designing and running autonomous clinical trials, are more speculative, he added.
Effective and ethical implementation of AI emerged as a key theme throughout the conference, held May 6-7 in Barcelona, Spain.
Strategies for effectively integrating AI into the workplace
AI offers an opportunity to streamline and accelerate clinical research work processes that are often hampered by inefficiencies, Mashlak said. “We are addicted to science and hungry for executions,” he added. Nevertheless, he says small increases in daily activity can have a big impact.
Currently, the best use cases for AI integration are high-intensity, high-reproducibility workflows, Mashraq says. Companies should “start small” with simple, low-risk processes that provide quick returns, he added. He warned against the so-called “pilot trap,” where companies tend to launch large numbers of disconnected pilots, such as AI chatbots, without an effective operating model.
He explains that such pilots often fail due to weak workforce planning, change management, and lax governance. To illustrate this, he cited a McKinsey survey of life sciences companies that claim to use generative AI, but only 32% of them took steps to scale the technology, and just 5% saw it as a key differentiator driving value.
Speakers on another panel at the conference emphasized the importance of training employees to use AI-based tools. “We have the solutions, but at the same time we have to learn how to use them,” says Kamil Sitters, COO and member of the executive committee of oncology biotech company Ryvu Therapeutics.
Despite its importance, AI-based training can be challenging from a resource standpoint, especially for small and medium-sized businesses, Sitters says. Another McKinsey report suggests that training employees to use new technology costs five times more than the tools themselves, Mashlach said.
The perfect workflow for early AI adoption
Mashraq says the strongest evidence for the use of AI is in applications for participant recruitment. He added that matching systems such as TrialGPT have reduced screening times by up to 42%, while OncoLLM has reduced review times to 3 to 12 minutes and increased accruals by up to 39%.
At AstraZeneca, Mashlaq and his team are integrating AI into simple, repetitive procurement tasks in vendor selection to streamline the process and reduce burden. This includes using AI to analyze structured data such as cost-based information. This is easier to compare and is primarily used in low-risk contracts. Additionally, AI can identify missing information in proposals based on a list of criteria, speeding up requests for additional information and reducing back-and-forth, he added.
Meanwhile, Alligator Biosciences, a small biotech company based in Lund, Sweden, is likely to buy AI solutions from vendors rather than build them in-house, said Karin Nordblad, the company’s director of immuno-oncology clinical operations. Alligator is still in the early stages of integrating AI into its daily operations, with some technically skilled employees pioneering it internally and sharing what they learn with others, Nordblad said.
The importance of transparency and consent
According to a 2025 survey conducted by the Center for Information and Research on Clinical Research Participation (CISCRP), 75% of 12,887 respondents said they were “somewhat” or “very” comfortable with AI being used to analyze medical data. However, 89% said they believe it is “somewhat” or “very” important to disclose the use of AI in clinical research.
Behtash Bahador, senior director of community engagement and partnerships at CISCRP, said these findings are particularly important as trust in pharmaceuticals remains low and lags behind other organizations involved in clinical research. Only 18% of respondents said they had a great deal of trust in pharmaceutical companies conducting clinical research, compared to 43% for government research agencies. Given these findings, “the last thing you want to do is to further erode trust by not disclosing your use of AI,” Bahador says.
Patient consent to participate in clinical research includes a detailed and accessible explanation of exactly how the AI will be used, says Branka Hezerova, associate director at GSK. Some patients may not want their data to be analyzed by AI, she added. Additionally, AI is known to introduce algorithmic bias, and not all participants will experience AI-supported tools in the same way. Therefore, patients who perceive themselves to be at higher risk of bias, such as those from minority groups, may be less likely to consent to research involving AI, which could reinforce existing inequalities, she explains.
Mashlaq said that under AstraZeneca’s AI-assisted vendor selection program, a comprehensive agreement details what data will be processed by AI and how that processing will take place. The transparency of this process alleviates potential vendor concerns about data security, he added.
Interested in joining or sponsoring OCT Europe? Please fill out the form below. The Arena team will contact you.
