PSI Reduces Clinical Trial Site Identification From Weeks to Minutes Using AI Agent Powered by Arango Contextual Data Platform

Clinical research organization uses AI agent-driven knowledge platform built on a unified contextual data layer to identify higher-performing trial sites faster and minimize inclusion of costly non-enrolling institutions

SAN JOSE, Calif., March 19, 2026 /PRNewswire/ -- NVIDIA GTC — Arango today announced that PSI CRO, a global clinical research organization, has reduced clinical trial site identification from up to six weeks to minutes using SYNETIC™, an AI-enabled knowledge engine powered by the Arango Contextual Data Platform. By unifying fragmented clinical research data into a trusted contextual data layer, PSI enables researchers to identify higher-performing trial sites faster, reduce non-enrolling institutions, and potentially save millions of dollars per clinical trial.

Selecting the right clinical trial sites is one of the most critical and expensive decisions in drug development. Clinical trials can take 10-15 years to bring a drug to market, and operational costs can exceed $160 per minute. Yet the industry faces a persistent inefficiency: 30-40% of clinical trial sites under-enroll and ~15% never enroll a single patient.

Activating a trial site can cost $30,000 or more, and large trials often involve hundreds of sites. When multiple sites fail to recruit patients, the financial impact can quickly reach millions of dollars per study.

The challenge is not a lack of data. Clinical research organizations already collect extensive information about investigators, institutions, and past studies. However, this critical knowledge is often fragmented across multiple systems, making it difficult to understand how investigators, institutions, protocols, and outcomes relate to one another.

Building a Contextual Foundation for AI in Clinical Research

To address this challenge, PSI built SYNETIC™, an AI agent-enabled clinical research knowledge base powered by the Arango Contextual Data Platform.

Using Arango, PSI unified structured data, documents, and historical clinical trial information into a single contextual data layer that preserves relationships across investigators, institutions, study protocols, patient populations, and historical outcomes.

By combining graph relationships, vector embeddings, documents, key value data, and search capabilities in a single multimodel platform, Arango enables PSI to analyze how clinical research data connects across hundreds of thousands of historical projects.

This unified, current, and trusted business context allows researchers to detect patterns in historical studies, identify investigators and institutions most likely to recruit patients successfully, and generate site recommendations in minutes rather than weeks.

Explainable AI for Regulated Environments

In clinical research, decisions must be transparent and defensible. AI systems must not only generate recommendations but also explain the reasoning behind them.

SYNETIC™ provides explainable insights for site selection decisions, including:

• the rationale behind site recommendations
• supporting evidence drawn from historical trial data
• confidence levels for predictions
• visibility into missing information or knowledge gaps

This transparency enables PSI researchers to trust AI-generated insights while maintaining the accountability required in regulated healthcare environments.

Transforming Clinical Trial Site Selection

Before SYNETIC™, compiling a final list of clinical trial sites could take up to six weeks of manual research and coordination across multiple systems.

Today, PSI teams can generate data-driven site recommendations in minutes. By identifying institutions more likely to recruit patients successfully, PSI can reduce the number of non-enrolling sites in a study – helping accelerate trial timelines and avoid millions of dollars in unnecessary costs.

"For AI agents to be useful, teams need to trust the recommendations," said Andrei Seryi, Director of Knowledge Management and Process Improvement at PSI CRO. "Our AI agent doesn't just recommend trial sites – it explains the reasoning, highlights key study factors, and shows confidence levels, all grounded in a unified, current, and trusted business context."

"Clinical trials depend on understanding complex relationships across enormous amounts of data, but too often that context is fragmented across systems," Seryi added. "With the Arango Contextual Data Platform, we can unify that information, ground AI in trusted and explainable data, and identify the right study sites faster – helping avoid underperforming sites that can add millions to the cost of a single trial."

A New Foundation for AI in Clinical Research

PSI's experience highlights a growing reality across many industries: successful AI initiatives depend not only on models and algorithms, but on the contextual data infrastructure that supports them.

The Arango Contextual Data Platform enables organizations to connect fragmented enterprise data across siloed systems while preserving the relationships between people, institutions, processes, and outcomes. By combining graph, vector, document, key value, and search capabilities in a single multimodel platform, Arango provides the unified contextual data foundation needed to power explainable AI and data-driven decision-making at scale.

"One of the biggest barriers to deploying AI agents in complex industries is fragmented data and missing business context," said Shekhar Iyer, CEO of Arango. "Too often organizations try to stitch together fragmented databases and pipelines to support AI, but those architectures were never designed to represent how enterprise data connects. By building SYNETIC™ on a unified contextual data layer, Andrei Seryi and his team at PSI created a trusted foundation for AI that enables faster and more confident clinical trial site selection."

About PSI CRO

PSI is a global full-service clinical research organization (CRO) dedicated to delivering quality and on-time clinical research services across numerous therapeutic areas. PSI supports pharmaceutical and biotechnology companies worldwide with end-to-end clinical trial conduct.

About Arango
Arango delivers a unified, natively multimodel contextual data platform that powers AI agents, assistants, and applications with the unified, current, and trusted business context needed to reason, decide, and act at scale.

The Arango Contextual Data Platform connects fragmented enterprise data with LLMs, copilots, and AI agents through a simplified architecture delivered out of the box. By combining graph, vector, document, key-value, and search capabilities in a single platform, Arango eliminates the complex stacks many organizations build to operationalize enterprise AI.

Trusted by organizations including NVIDIA, HPE, the London Stock Exchange, PSI CRO, the U.S. Air Force, NIH, Siemens, Transient.AI, Matpriskollen, and Articul8, Arango helps enterprises move from AI pilots to reliable production systems faster while lowering infrastructure complexity and total cost of ownership. Arango is a proud member of the NVIDIA Inception Program and the AWS ISV Accelerate Program. Learn more at arango.ai

Media Contact
[email protected]

SOURCE Arango