A new global study of 285 data, AI, IT, and business stakeholders finds that organizations with mature context engineering programs are four times more likely to qualify as AI leaders than their peers. The study, “Context Engineering for Agentic AI: Architecture, Use Cases, and Principles for Success,” published by BARC (Business Application Research Center) and sponsored by DataHub, is available as a free download here.

The BARC study classifies 42 percent of respondents as “context leaders,” organizations that have implemented, formalized, or optimized six foundational elements of context engineering, including data integration, workflow orchestration, retrieval methods, federated metadata, prompt engineering, and the semantic layer. Among context leaders, 49 percent also qualify as AI leaders, compared with 12 percent of all other organizations surveyed.

The research surfaces a gap between how enterprises build context for individual AI agents and how they govern context across the organization. Forty-four percent of respondents manage context within a single agent, team, or platform, while 43 percent manage it across teams, platforms, or the full enterprise. Organizations in the latter group are better positioned to reuse context across business domains, improving both accuracy and efficiency as they scale agentic AI deployments.

“Agentic AI fails without business context. Agents can turn an inaccurate answer into a bad decision or action,” says Kevin Petrie, VP of Research at BARC US and co-author of the study. “Organizations that establish shared meaning, controlled retrieval and governed memory give their agents a stronger foundation for reliable and auditable work.”

Thirty-eight percent of respondents cited consistency and reliability as their top priority for context engineering, while 34 percent cited accuracy. Cost and effort reduction ranked lower at 12 percent, though BARC expects that priority to rise as organizations scale token consumption.

Data quality and preparation ranked as the leading challenge, cited by 49 percent of respondents, followed by AI model limitations at 29 percent and governance gaps at 25 percent. Data freshness, at 23 percent, reflects the ongoing difficulty of keeping context current as source data and business conditions change.

“Context engineering is the practice of assembling the right inputs for a single AI agent call, while context management is the enterprise discipline of governing those inputs across many agents and data sources at scale,” said Shirshanka Das, co-founder and CTO of DataHub, which sponsored the report. “This research confirms what we see with our customers. The organizations pulling ahead on AI treat context as shared infrastructure rather than a series of one-off engineering projects.”

The full BARC study, including architectural frameworks, adoption benchmarks, and a practical guide for data and AI leaders, is available here.

About DataHub

DataHub transforms enterprise data into trusted context, enabling intelligent decision-making by humans and AI agents. The company was founded by the creators of the popular DataHub open source product that has more than 16,000 contributors and is used by thousands of organizations. The company’s flagship product, DataHub Cloud, is the leading context management platform trusted by the Global 2000 to ensure that context is always relevant, reliable, and continuously refreshed across the entire data estate. DataHub is backed by Bessemer Venture Partners, LinkedIn, and 8VC. For more information, go to: https://datahub.com/.

Media gallery

About The Author
Business Wire