AstraZeneca is emerging as a frontrunner in the pharmaceutical industry’s race to apply artificial intelligence to clinical trials, not by focusing solely on internal research gains, but by embedding AI directly into national healthcare systems and everyday patient care.
Across Big Pharma, AI is reshaping drug discovery, development, and trial design. Many companies use machine learning to speed up molecule identification or streamline internal processes. AstraZeneca’s approach stands out because its AI tools are already operating at public health scale, screening hundreds of thousands of people and influencing how care is delivered in real-world settings.
That distinction is backed by clinical evidence. At the European Lung Cancer Congress in March 2025, AstraZeneca presented results from its CREATE study, which evaluated an AI-powered chest X-ray screening tool. The system achieved a positive predictive value of 54.1%, well above the predefined success benchmark of 20%. These are not theoretical gains. Since 2022, more than 660,000 people in Thailand have been screened using the technology, with suspected pulmonary lesions identified in roughly 8% of cases.
The program has moved beyond pilot status. Thailand’s National Health Security Office is now expanding the AI screening initiative across 887 hospitals, supported by a three-year budget exceeding 415 million baht. This makes it one of the clearest examples to date of AI-driven clinical trial technology being deployed across an entire national healthcare system.
Different paths in Big Pharma’s AI race
Other major drugmakers are also investing heavily in AI, but with different priorities. Pfizer, for example, has focused on accelerating drug discovery. Its machine learning research hub has reduced molecule identification timelines to about 30 days, and AI played a role in the rapid development of its COVID-19 treatment, Paxlovid. Pfizer now uses AI in more than half of its clinical trials, largely to increase speed and efficiency within its pipeline.
Novartis has taken a similar approach through partnerships with Isomorphic Labs, founded by Nobel laureate Demis Hassabis, and Microsoft. Its AI-driven systems use “computational twins” to simulate trial processes, helping the company identify trial sites that can recruit patients faster than traditional methods.
Roche, meanwhile, has built an integrated data strategy through acquisitions such as Foundation Medicine and Flatiron Health. By combining AI models with laboratory experimentation and one of the world’s largest clinical genomic databases, Roche aims to improve safety monitoring efficiency by up to 50% by 2026.
AstraZeneca’s operational focus
What differentiates AstraZeneca is its emphasis on clinical operations rather than discovery alone. The company is running more than 240 global trials and has introduced generative AI across multiple stages of trial execution.
Internally, an AI-assisted protocol development tool has cut document authoring time by as much as 85% in some cases. Imaging tools now help identify and map locations on CT scans, significantly reducing the time radiologists spend on manual annotation.
More fundamentally, AstraZeneca is experimenting with virtual control groups. By using electronic health records and historical trial data to simulate placebo arms, the company aims to reduce the number of patients who receive non-active treatments. This approach could reshape how trials are designed, particularly in areas where recruiting participants is difficult or ethically complex.
The lung cancer screening program in Thailand illustrates this broader strategy. Using Qure.ai’s qXR-LNMS tool, AstraZeneca is not just collecting data for trials but strengthening local healthcare infrastructure. An expansion planned for late 2025 includes screening industrial workers across four Thai provinces and extending AI detection beyond lung cancer to conditions such as heart failure.
Why speed and scale matter
The stakes are high. Traditional drug development typically takes 10 to 15 years, with failure rates approaching 90%. Early data suggest that AI-assisted drug candidates may achieve Phase I success rates of 80% to 90%, compared with 40% to 65% for conventional approaches. More than 3,000 AI-supported drugs are currently in development, and analysts expect over 200 AI-enabled approvals by 2030.
While companies like Pfizer and Novartis focus on shaving months or years off development timelines, AstraZeneca’s model delivers immediate benefits to patients, particularly in underserved populations. Its AI tools are detecting disease earlier, sometimes before symptoms appear, while also generating real-world evidence under regulatory oversight.
The bigger economic question
The World Economic Forum estimates that AI could generate between US$350 billion and US$410 billion annually for the pharmaceutical industry by 2030. The open question is where that value will concentrate. Faster discovery, smarter trial design, and integrated diagnostics all promise returns.
AstraZeneca is betting that embedding AI throughout clinical operations, from protocol writing to patient recruitment and regulatory submission, will deliver both speed and credibility. Rather than acquiring AI startups outright, the company has favored partnerships with technology firms, regulators, and national health systems, particularly in regions where infrastructure gaps exist.
As AstraZeneca works toward its goal of delivering 20 new medicines and reaching US$80 billion in annual revenue by 2030, its AI strategy is less about headline-grabbing algorithms and more about proof under real-world conditions.
