GoodData.AI Launches AI Observability to Track, Trace, and Trust Enterprise AI

AI Observability connects AI adoption, quality and cost with interaction-level tracing and intelligence, helping teams understand what happened at every step and what to improve next.

SAN FRANCISCO, California - [23.09.2026] - GoodData.AI today announced AI Observability, an extension of its agentic infrastructure that gives enterprises visibility into how AI is being used, how it is performing, what it costs, and why individual responses behave the way they do.

Traditional monitoring shows when AI usage, quality, or cost changes. It rarely shows why. GoodData.AI AI Observability connects organization-wide performance signals with the execution path behind individual interactions, so teams can move from spotting a problem to fixing its cause.

GoodData.AI provides a continuous path from usage analytics to interaction tracing to intelligence. Teams can see which AI experiences are being adopted, inspect the skills, knowledge, memory and model calls behind a specific response, identify failure points, and surface recurring issues across conversations.

From AI performance to the reason behind every answer

AI Observability brings three levels of visibility into one experience:

Understand AI usage at scale. Teams can track query volume, active users and workspaces, adoption across agents and skills, quality signals, token consumption and cost. This gives product, data and AI leaders an organization-wide view of where AI is being used and where investment is delivering value.

Trace individual interactions. When a response needs investigation, teams can inspect its execution path step by step, including which skills were considered and activated, what knowledge and memory were retrieved, model calls, failures, timing, iteration counts and token usage.

Turn interactions into intelligence. GoodData.AI can analyze conversations for recurring issues, surface patterns across interaction history and recommend changes to knowledge, semantic models or configuration, helping teams improve AI quality over time.

The architecture follows the model GoodData.AI developed internally: aggregate usage analytics establish whether AI is being used and working; interaction observability explains what happened inside a response; and interaction intelligence identifies what to change to make results better.
"Enterprise teams need more than a dashboard telling them that AI quality or cost changed. They need to understand what produced an answer: which skills ran, what knowledge and memory were used, where execution failed, and what they should change next. AI Observability connects those layers so teams can improve AI with evidence rather than intuition."

Rosta Striz, Principal Product Manager at GoodData.AI

Observability without another data pipeline

AI Observability runs as a managed workspace using interaction data the AI already generates. Organizations can begin with prebuilt dashboards and customize their own views on the same data, without creating a separate observability pipeline. It supports deployment in GoodData.AI's cloud or on customer-controlled infrastructure.
This keeps observability close to the analytics environment itself. GoodData.AI provisions observability directly within the customer organization, with prebuilt metrics and dashboards and the ability to build additional analytics on the same underlying interaction data.

Built for every team responsible for AI
For AI engineering teams, interaction tracing provides evidence for debugging agent behavior and improving reliability. For data and analytics teams, it shows where data, definitions, knowledge or semantic logic affect AI quality. Product teams can compare adoption and engagement to decide which AI experiences warrant further investment, while compliance and governance teams gain traceability and audit history for reviewing AI behavior.

The result is a common observability layer across the teams that build, operate and govern enterprise AI.

How the pieces fit together

AI Observability extends GoodData.AI's broader approach to Agentic Serving Plane. The company's Context Management framework already brings together governed semantics, enterprise knowledge, guidance, governance and observability to help organizations operate AI reliably in production.

With AI Observability, that transparency becomes operational: enterprises can measure how AI is being used, investigate how an individual answer was produced, and use patterns across real interactions to improve the system.

"AI becomes much more valuable when enterprises can see how it works and improve it continuously. Observability turns every interaction into evidence: evidence about adoption, quality, cost and the context behind the answer. That gives enterprises the control they need to scale AI with confidence."

Roman Stanek, CEO and Founder at GoodData.AI

About GoodData.AI

GoodData.AI is an open agentic analytics platform that lets enterprises put AI to work on their data without losing control of it. Its agents follow through on a business process rather than answering one question and stopping. Every answer and action they take is grounded in business definitions, permissions, and context the customer owns, so the people using it, not the AI, decide what happens next.

The same governed platform lets enterprises build and run many such agents across the business without redoing the governance work each time. Each one can be improved rather than replaced as needs change.

The platform supports customer-controlled infrastructure, bring-your-own-LLM flexibility, MCP and A2A integration, and open development through APIs and SDKs. Headquartered in San Francisco with engineering based in Prague, GoodData.AI serves enterprises and software companies worldwide.

For more information, visit gooddata.ai and follow GoodData.AI on LinkedIn.

GoodData.AI Contact press@gooddata.com

© 2026 GoodData Corporation. All rights reserved. GoodData.AI is a trademark of GoodData Corporation.