OpenText and Google are pushing to close a persistent data-layer gap by turning decades of ungoverned, unstructured corporate information into the contextual inputs agentic AI needs to operate reliably.
- OpenText and Google spotlight context engineering at Google Cloud Next 2026
- Legacy, unstructured data is blocking reliable agentic AI deployments
- Enterprise success depends on curated, governed context as well as models
What happened
At Google Cloud Next 2026, OpenText and Google emphasized a shared concern: enterprises cannot rely on raw model capability alone to run agentic AI at scale. Instead, they highlighted the need to build a usable data layer that extracts, curates and governs context from long-standing information systems.
Their presentations and messaging framed ‘context engineering’ as an operational discipline aimed at transforming scattered, unstructured corporate content into actionable inputs for agentic agents.
Why it matters
Organizations that skip the data-layer work risk deploying agents that produce unreliable or noncompliant outcomes because models lack the accurate, current context needed for decisioning. Decades of legacy records, email, documents and other ungoverned sources create a hidden friction point that undermines the benefits of improved model architectures.
Addressing this gap is not just a technical challenge but a governance and adoption issue: IT leaders must balance data access, quality and control to unlock safe, scalable agentic workflows across the enterprise.
What to watch next
Expect momentum around tooling and partnerships that promise to surface, standardize and govern contextual data for agentic AI. Watch for tighter integrations between content-management vendors and cloud platform providers as companies try to operationalize context engineering.
Also monitor enterprise buying signals: procurement and architecture decisions will reveal whether organizations prioritize context work alongside model procurement, and how quickly governance practices evolve to support agentic deployments.