Overview
Today we're launching AI Analyst and the Daasity MCP Server to all merchants. Both let you ask questions about your business in plain English and get answers built from the same governed metric definitions that power your dashboards.
AI Analyst lives inside Daasity. Ask "which products drove growth last month" or "what changed last week, and why," and you get an answer with its assumptions shown, suggested follow-ups, and a View data option that expands the underlying rows.
The MCP Server connects the AI tools your team already uses — Claude, Cursor, Codex, ChatGPT — directly to your Daasity data. One connection, and your AI answers from live, governed numbers instead of whatever spreadsheet was last exported. Build a custom report from a prompt, combine a retailer's category review with your own margins, or schedule a Monday recap that writes itself.
Same engine, two doors. The distance between a business question and an answer you can act on just got a lot shorter.
What we changed
Both surfaces resolve questions against your account's governed semantic views — the metric definitions generated from your own data model, not raw warehouse tables. A question is matched against your dashboards first, then your deployed semantic views, and only falls back to generated SQL when neither covers it, in which case the answer says so.
The MCP Server is self-serve. Open the MCP screen in your Daasity profile menu, copy your connection URL, and add it as a custom connector in your AI tool. Every connection is read-only, scoped to your workspace, expires after 30 days, and can be revoked by you at any time.
Supported on Snowflake, BigQuery, and Redshift. Looker is not required — merchants without a BI tool are provisioned automatically on first login.
What to expect
AI Analyst and MCP access are enabled per user. If you'd like your team turned on, send your account manager the names and emails, and for multi-brand organizations, which brands each person needs. Turnaround is same-day during this period.
Setup takes minutes and the step-by-step guides are live in the Help Center:
A good first question: "Top ten products by net sales in the last 30 days, with units." Then open the equivalent dashboard and compare. That's what builds trust fastest.