Accelerating Clinical Trials with AI: The Keck School of Medicine of USC Story

Part of Pluto7’s Decade of Impact

Clinical trials sit at the heart of medical innovation. Every improvement in the speed and efficiency of these trials can have a direct impact on how quickly life-saving therapies reach patients.

At the Keck School of Medicine of the University of Southern California (USC), researchers and administrators faced a common challenge shared by many research institutions: clinical trial administration was highly manual, time-consuming, and difficult to scale.

To address this, the Keck School partnered with Pluto7 and Google Cloud to bring machine learning into clinical trial management—transforming how key administrative processes are executed and dramatically accelerating trial activation timelines.

The Challenge: Administrative Bottlenecks in Clinical Trials

Clinical trials require extensive coordination across regulatory, financial, and operational teams. One of the most complex and time-intensive processes is Medicare Coverage Analysis (MCA).

MCA determines whether each procedure in a clinical trial should be billed to the research sponsor or covered by insurance as part of standard care. Traditionally, this process requires experienced administrators to manually review clinical procedures and determine the appropriate billing category.

For the Keck School’s Clinical Trials Office, this meant:

• Manual review of complex study procedures
  
• Administrative bottlenecks slowing down trial activation
  
• Limited access to real-time data across systems
  
• Significant staff time required to manage approvals and budgeting
  

These inefficiencies could delay the launch of important research studies and slow down the broader pace of medical innovation.

The Keck School knew that advanced analytics and machine learning could help modernize these processes.

Building the Foundation for Advanced Analytics

The transformation began with a deep review of clinical trial administration workflows.

Working with process improvement experts, the university streamlined its internal operations and eliminated 30% of the steps required to approve clinical trials, while reducing activation times by 50%.

With these improvements in place, the next step was enabling data-driven automation.

To achieve this, the Keck School turned to Google Cloud’s advanced analytics platform, combined with Pluto7’s expertise in machine learning and AI-driven workflows, powered by Planning in a Box – Pi Agent.

Together, the teams built a scalable analytics environment capable of supporting real-time data access, machine learning models, and automated workflows for clinical trial management.

Applying Machine Learning to Streamline Clinical Trial Administration

Pluto7 worked closely with USC researchers and administrators to understand the most time-consuming processes within clinical trial operations.

One key opportunity was automating the Medicare Coverage Analysis workflow.

Using machine learning models deployed on Google Cloud, Pluto7 developed an algorithm capable of:

• Reading standard-of-care clinical guidelines
  
• Matching trial procedures with those guidelines
  
• Automatically assigning the correct billing category
  

What once required manual review over several days could now be completed in milliseconds.

Google Cloud technologies such as Document AI, Cloud Vision, and BigQuery helped accelerate data processing and enable real-time analytics.

The result was a highly collaborative solution where domain experts worked closely with data scientists to train and refine the algorithms for real-world clinical trial workflows.

Transforming Clinical Trial Speed and Efficiency

The new ML-driven system significantly improved the speed and efficiency of clinical trial administration.

Key outcomes included:

• Medicare Coverage Analysis reduced from days to milliseconds
  
• Significant reduction in clinical trial activation time
  
• Improved efficiency of clinical trial administration
  
• More time and resources freed for research teams
  

The machine learning model also demonstrated 70–90% prediction accuracy in identifying the correct billing designation, an important improvement over manual workflows.

With approximately 200 clinical trials conducted annually at USC, even modest time savings translate into major efficiency gains across the organization.

A Foundation for Future Medical Innovation

The success of the initial implementation has opened the door for further innovation.

Today, the Keck School continues to work with Pluto7 to expand machine learning capabilities across additional research workflows. By leveraging BigQuery and serverless analytics infrastructure, the institution is exploring ways to combine internal research data with large public datasets to unlock deeper insights.

These capabilities will enable faster analysis of large clinical datasets and support more advanced research into population health and clinical outcomes.

For the Keck School, the goal is clear: accelerate the pace of discovery while improving patient outcomes.

The Impact

By combining Google Cloud’s advanced analytics platform with Pluto7’s Planning in a Box – Pi Agent, the Keck School of Medicine established a powerful foundation for modern clinical research operations.

The collaboration demonstrates how AI and machine learning can transform complex administrative processes, freeing researchers to focus on what matters most: advancing medical science and improving patient care.

As clinical research continues to evolve, innovations like these will play a critical role in helping institutions bring new therapies to patients faster.

Explore how AI and Planning in a Box – Pi Agent can accelerate data-driven decision making across complex workflows.