Two new ways to get answers in Bigeye, for people and AI agents

Two kinds of questions come into Bigeye every day. The first is from an analyst or data steward who wants to know or do something: is this table fresh, why did this metric drop, which pipelines downstream of this source are affected, or create a new data quality check like a null check for user_id. Answering it means navigating to the right view, applying the right filters, and interpreting what's there. Workable for experienced users, friction for everyone else.

The second kind of question comes from an AI agent. A team running Claude Code, GitHub Copilot, or a custom agent hits a point where the agent needs to know something about data quality or lineage before it can proceed. The usual answer is a hand-rolled script calling the Bigeye REST API, or a manual step where someone pulls the relevant information and pastes it into the agent's context.

These are different problems with a common pattern: getting answers out of Bigeye requires more steps than it should. Today we're introducing two additions to the AI Trust Platform that address both.

bigAI Chat: ask Bigeye a question, get an answer with context

bigAI Chat is a conversational interface built into Bigeye. Type a question in plain language ("which tables feeding this dashboard have open quality issues," "show me the lineage for this column," "why did volume drop on this pipeline overnight," or "create a freshness check for the daily sales table") and bigAI Chat returns an answer with context from your environment.

bigAI Chat surfaces answers and proposes next steps. A human approves before anything changes. It works alongside your SQL and BI tools, not as a substitute for them. For analysts onboarding to a new environment, asking a question is faster than learning where in the platform to look for it. For experienced users, it compresses what used to be five minutes of clicking into thirty seconds.

A note on naming, because precision matters here: bigAI Chat is an extension of bigAI, the existing feature that runs automated root-cause analysis and surfaces incident recommendations. bigAI works in the background on your behalf. bigAI Chat responds to your questions. Both are part of the platform; they're doing different things.

MCP Server: governed access for AI agents, using permissions already in place

MCP Server gives AI agents access to Bigeye through the Model Context Protocol, the open standard for connecting AI agents to external tools and data sources. An agent running Claude Code, Snowflake Cortex Code, or GitHub Copilot CLI can check data quality issues, trace lineage, search the catalog, profile tables, and manage monitors and dimensions, all without a human copying results from the Bigeye UI and pasting them into the conversation.

Access runs through the same permissions model already governing your Bigeye environment. An agent operating through MCP Server sees what its credentials allow and nothing else. There's no separate access layer to configure, and no new integration to maintain when permissions change.

For teams who've been writing one-off scripts to pull Bigeye data into agent workflows, MCP Server replaces that with a standard every compliant AI client already speaks. For teams evaluating whether to give AI agents access to Bigeye at all, the permissions and audit record questions have a concrete answer: same model, same log.

Teams that need to self-host can. MCP Server is open source, and the documentation covers setup for each supported client.

The same data, two front doors

bigAI Chat is the human-facing front door. MCP Server is the machine-facing front door. Both connect to the same Bigeye environment, with the same catalog, quality signals, and lineage, and the same permissions model governing access in both directions.

The Agent Trust Hub operates at the governance layer for both. Conversations in bigAI Chat and agent queries through MCP Server both generate an audit record: what was asked, which data was accessed, and what quality or classification signals that data carried. For data and compliance teams responsible for what AI touches in their environment, that record is how you see it.

How to get started

bigAI Chat is available now for all Bigeye customers. You'll find it in the platform with no separate setup required. To get started with MCP server, start here.