Elastic, the Search AI Company, enables everyone to find the answers they need in real time, using all their data at scale. Internally, the data organization is applying that same philosophy to how teams discover, trust, and use data—focusing on building a strong, governed foundation of structured, well-documented, and context-rich data as the prerequisite for any future AI-driven analytics.
Why start in Finance
Finance is where accuracy, accountability, and trust matter most. The demand for fast, reliable answers is constant, yet even small gaps in definition or logic can quickly erode trust. Leaders want timely insight without waiting on engineers for every iteration, but those answers must be precise and defensible.
The Finance domain was selected as the first proving ground because it reflects the overall health of the company and carries the highest expectations for accuracy and auditability. It is both high-impact and highly measurable—making it the ideal place to demonstrate how strong data foundations can directly improve decision-making before expanding to other domains.
Building the Documentation Layer
At Elastic, documentation isn’t about checking a box — it’s about creating rich, meaningful data narratives that both people and AI can truly understand. Simple field descriptions aren’t enough when the goal is to enable confident decisions, automated insights, and future AI-driven analytics. What’s needed is context-aware documentation that explains not just what the data is, but why it exists, how it’s created, and how it should be trusted and used.
Previously, Elastic’s definitions were fragmented across dbt YAML files, Atlan, and other tools. As teams moved at different speeds in their analytics journey, documentation quality varied widely, leaving critical gaps in shared understanding.
To modernize this process, Elastic is leveraging AI-driven documentation through Elementary’s MCP Server to automate and scale documentation across the data platform. These tools don’t just fill in missing text; they help generate deep, text-rich, business-aware documentation, propose meaningful model and column descriptions, and automatically propagate consistent definitions across lineage. The result is a living metadata layer that scales with the data—and forms a trusted foundation that both human users and AI systems can rely on with confidence.
Developer workflow
Elastic is prioritizing AI-driven documentation and testing as a core part of the development workflow from day one. As new models are created in dbt, Elementary’s MCP Server and AI agents actively drive the creation of foundational documentation at the point of development—ensuring YAML files are generated, model and column descriptions are automatically proposed, and baseline tests such as not null and unique tests are automatically applied.
Analytics engineers own the initial technical documentation and data quality standards, while business users and the governance team can later enrich the models with business context. By leveraging MCP to continuously infer context from source applications, Elastic is shifting documentation from a manual afterthought to an AI-assisted, real-time process that scales with development.
Bringing Data Quality Context into Atlan
Elastic uses Atlan as the central system for data discovery, documentation and governance, with Elementary providing real-time observability, testing, and lineage. When connected, the two platforms unify metadata, data quality signals, and trust context into a single experience.
For business users, this means they can remain in Atlan and instantly understand not only what a dataset represents, but also whether it is healthy, under investigation, or impacted by an active issue—without switching tools. By surfacing real-time reliability signals alongside rich documentation, Atlan and Elementary together provide a shared, authoritative view of data trust for both engineers and analysts, directly where decisions are made.
A focused pilot
The Finance domain provides the right combination of impact and clarity, centered on critical reporting models that directly inform executive decision-making. It serves as the ideal environment to establish the standards for trusted data, rich context, and AI-ready documentation.
This pilot is designed to become the playbook for how Elastic defines semantic context, documentation, and data quality to unlock AI-driven insights across the organization. Success is measured not by the number of dashboards produced, but by behavior change: developers naturally creating YAMLs and tests as part of development, engineers of all experience levels contributing meaningful documentation, and Finance users confidently interpreting reports without constant clarification.
Next steps
With the Finance pilot underway, Elastic will expand this approach to additional domains using the same repeatable playbook: automated, AI-assisted documentation from day one, embedded data quality testing in development, and real-time trust signals surfaced where business users work.
Over time, this consistent foundation will enable Elastic to safely scale AI-driven and natural language analytics, powered by the same rich semantic context that supports human decision-making today.
