The signal change in the past couple of years is the transition of data intelligence from primarily serving analytics and GenAI use cases in the enterprise toward emerging agentic AI use. This is something we noted in our own primary research, the 2026 CDO Insights Survey from early this year, for example, which indicates a lift-off in GenAI adoption, moving from 48% to 69% year-over-year.

For companies focused on harnessing agentic transformation, this shift represents a sea-change for how they approach data intelligence. Some of the key imperatives that companies can’t afford to fall behind on include:

1. Metadata that agents can actually read: Siloed human-centric data catalogs just can’t keep up with the demands of an agentic enterprise. Metadata can no longer exist as simple static documentation relying on human stewards to decode, understand and enrich with context before it’s delivered to AI agents. Metadata needs to be programmatically, updated and structured so that autonomous models, AI agents, and orchestration engines can consume context automatically.
2. Data products that are ready for AI, not just humans: To gain complete and reliable context of data, autonomous agents and advanced models cannot consume raw, uncurated tables without risking hallucinations or compliance failures. Agentic execution demands packaged, semantically enriched, and pre-governed "data products" that are ready for immediate AI consumption.
3. AI that helps manage the data itself: Data sprawl across multicloud environments makes manual tagging, lineage tracing, and policy enforcement unscalable. Organizations urgently require AI to scale data curation, automate routine governance workflows, and manage compliance at cloud scale. This shift from labor-intensive, manual governance to automated, AI-driven metadata management is critical for dealing with data management scale and speed that AI operations require.

Organizations can struggle to transition their AI agents into production. Time and again, we’ve seen that without adequate context and quality of data, AI agents can easily go off the rails. We believe this is where data intelligence plays a significant role. It enables practicable, production-ready agentic AI by improving the AI-readiness of data.

IDMC delivers high-quality, relevant, governed context, available directly in the agent runtime when and where it’s needed, through MCP server tools, agent-to-agent (A2A) interfaces, or public APIs. It also ensures the quality and reliability of that context through comprehensive, autonomous or expert-guided curation and stewardship, through its inventory of CLAIRE® Agents, which facilitate discovery, curation, quality, mastering and delivery across the data estate. Both sides of the metadata ledger, production or consumption, are covered by agentic, AI-enabled process flows for rapid, efficient value delivery now, and in the future.

This recursive dynamic, where agent-curated metadata context is made broadly available across agentic AI workloads, has been extensively tested at Salesforce, across many runtime AI harness environments. What we repeatedly note is that governed context pays continual dividends.

A recent study corroborates this, revealing that agents drawing on governed metadata used up to 29.8x fewer tokens than ungoverned approaches and returned more accurate results across every query and scale tier. Even as the test estate scaled 50x, governed token utilization barely budged, growing less than 50%.¹

This combination of AI utilization efficiency and agent task proficiency points the way forward for enterprise agentic AI: enabling data intelligence that delivers reliable, governed context agents unlocks more accurate and cheaper outcomes.

Our customers are seeing real benefits with this approach. Jotun is a century-old Norwegian coatings manufacturer operating 40 factories across five regions and serving customers in more than 100 countries. It used the migration of its legacy data platform as an opportunity to realize a platform-first modernization, setting themselves up for future agentic AI success.

Starting with Informatica’s Intelligent Data Management Cloud (IDMC) and its Cloud Data Governance and Catalog (CDGC) and CLAIRE services, they rapidly moved all their legacy data warehouse and analytics workloads to a modern, multi-cloud architecture that includes Snowflake, Azure, and Power BI.

Jotun achieved its data infrastructure modernization to the cloud in under six months by partnering with Informatica and Snowflake. The new, centralized data hub enables self-serve data catalogs to help local teams accelerate time to market. With 7x faster development, teams can streamline preparation for AI adoption and ESG compliance.

"Our partners at Microsoft and Snowflake assured us that Informatica is a huge player. Their advice was to think platform-first, instead of choosing disparate point solutions and then trying to stitch them together," says Gro Kamfjord, Head of Data at Jotun.

Informatica’s multi-cloud, catalog-of-catalogs supports a “metadata-as-a-service” approach, empowering regional teams and offices with data intelligence to adopt and consume as needed. The needed context is always somewhere, but data engineers didn’t necessarily know where to locate it or how to access it. With IDMC’s CDGC service, metadata scanners automatically find, curate, and track the lineage of data assets across apps like Power BI, Workday, SharePoint, and more. Pre-built cloud connectors, reusable templates and AI-powered CLAIRE with its recommended steps further helped speed up the development process.

This federation is what makes an agentic future practical: each affiliate can adopt AI at its own pace, informed by provider choice, cost-performance trade-offs, or local regulatory and policy considerations, but sourcing from a single, governed context foundation. Read more.