Is your data stack ready for AI?

The demands data teams are facing are changing faster than most data infrastructures were built to support. AI agents don't wait for batch jobs, don't tolerate inconsistent definitions, and don't forgive architectures designed for human-paced analytics. And yet, 85% of enterprises are running agentic AI on a data foundation that isn't ready to support it.

The gap between where most data stacks are and where agentic AI needs them to be isn't a tooling problem. It's an architectural one. That's what Open Data Infrastructure is designed to solve, and it's what the ODI Assessment is designed to measure and help fix.

What is Open Data Infrastructure?

Open Data Infrastructure (ODI) is an architectural approach built on open standards, interoperable storage, decoupled compute, and unified context. Its goal is to keep data portable, accessible, and governable across tools and environments, rather than binding access to a single platform's control plane.

While humans have tolerance for data latency and the experience and problem-solving skills for missing context, AI agents are a fundamentally different consumer. They operate continuously, autonomously, and at a scale of queries no human team could match. They require fresher data, consistent semantics across systems, and governed access paths that don't break under load.

ODI shifts the foundational architecture to meet those demands. It moves organizations from warehouse-centric, proprietary, analytics-driven stacks to lake-first, open-standards-based infrastructure that serves analytics, operations, and AI from the same source of truth.

The architectural components of an ODI

An Open Data Infrastructure is built on 3 interconnected pillars:

1. Automated, standards-based ingestion and transformation for architectural flexibility

Data ingestion and transformation logic should be portable and engine-agnostic. When pipelines and business logic are tightly coupled to a single compute vendor, any architectural change — a new warehouse, a new AI runtime, a new processing engine — requires rebuilding from scratch. Open standards like open connector frameworks for data movement and dbt for transformation keep that logic stable and reusable, so the stack can evolve without constant re-engineering.

2. An open data lake foundation to reduce overhead and eliminate duplication

Data should be stored once in object storage, in open formats (Apache Iceberg™, for example), making it readable by any compatible compute engine downstream. This decouples storage from compute, preventing the vendor lock-in that comes from proprietary storage formats and eliminating duplicate ingest fees that arise when the same data is synced to multiple destinations to serve varying use cases — especially important as AI query volumes scale. A unified lake layer means every downstream consumer (warehouse, AI agent, analytics tool) works from the same governed copy of the data, not a siloed version of it.

3. Unified activation, semantics, and AI consumption for trusted data across systems

Data is only as useful as the context around it. AI agents don't just need access to data; they need to understand what that data means. A semantic layer defines business logic (metrics, entities, relationships) once, and makes it consistently available to every system that consumes data: dashboards, models, pipelines, and AI agents. Without this layer, agents make assumptions about undefined terms and act on those assumptions at machine speed. With it, they query governed, pre-approved definitions and flag anything that doesn't exist rather than guessing.

Together, these 3 pillars create a data foundation that serves both human and machine consumers consistently, at scale, and without requiring constant re-architecture as AI tools evolve.

[CTA_MODULE]

Introducing the ODI Assessment

The ODI Assessment is a short, 8-question quiz designed to help teams understand where their current stack stands relative to an Open Data Infrastructure and what to prioritize to close the gap. 

The questions cover the fundamentals of your current architecture: where data lands, how it's ingested and transformed, what formats and platforms you're using, where you are in your AI journey, and how you're managing metadata and context for AI systems. It takes less than 5 minutes to complete. 

Take the assessment

See where your data stack stands and get a clear picture of what it would take to close the gap.

[CTA_MODULE]

Apache Iceberg is a trademark of the Apache Software Foundation.