ChatGPT on your data: ask in plain language, get your real numbers.
Connect the Polar MCP to ChatGPT and it reads your governed semantic layer directly, net sales, blended CAC, contribution margin, defined once and the same as your dashboards. No CSV, no copy-paste, and no number invented to fill a sentence. ChatGPT does the reasoning; Polar supplies the truth.
/Setup Guide
How to Connect ChatGPT to Polar.
A Polar account with your stack connected, ChatGPT with connectors enabled, and the Polar MCP. Then ask.
01 · What you need
A Polar account with your stack
A Polar account with Shopify and your sources connected. ChatGPT reads the same governed metrics your dashboards do.
Connect your data →ChatGPT with connectors
A ChatGPT plan that supports connectors, on claude.ai or the desktop app, so the Polar MCP can be added to your workspace.
Open ChatGPT →The Polar MCP
One connection. It exposes your metrics and dimensions to ChatGPT as a queryable layer, read-only, scoped to your workspace.
02 · Connect Polar to ChatGPT
1
Add the Polar MCP ~2 min
One connection, read-only.
In ChatGPT, add the Polar MCP as a connector and authorize your workspace. ChatGPT can now read your governed semantic layer,
net sales, blended CAC, contribution margin, without any export.2
Ask in plain language instant
No SQL, no dashboard hunting.
1
Ask the question the way you would ask an analyst: by channel, by cohort, by product, over any period.
2
ChatGPT resolves it against Polar's metric definitions, so the answer matches the dashboard rather than competing with it.
what was our blended CAC last week vs the week before
break contribution margin down by product for June
break contribution margin down by product for June
3
Reason, then act the payoff
Analysis, not just retrieval.
Because the numbers are trustworthy, the reasoning on top of them is too. Ask ChatGPT to find the cause, draft the weekly review, or turn the answer into a plan, and every figure it cites is sourced to Polar.
Side by side
An assistant without your data guesses. Polar makes it answer.
Ask a general assistant about your business and it works from whatever was pasted into the chat. Connected to Polar over MCP, ChatGPT reads your governed semantic layer instead, so the figures in the answer are the same ones on your dashboards.
ChatGPT on its ownon its own
+ChatGPT with Polarblended
Numbers come from whatever was pasted into the chat.
Metrics fetched live from your governed semantic layer.
An answer can look right and still be invented.
Every figure is retrieved, so there is nothing to hallucinate.
No view across Shopify, ads, email and the rest.
Sources already blended before the question is asked.
Yesterday's export is out of date by the time it is read.
Live data, on the same definitions as your dashboards.
The Gap
The connector is only half of it.
ChatGPT reads Polar's governed metrics over MCP, so the answer is fetched from your semantic layer rather than assembled from whatever was pasted into the chat. First-party, sourced, and the same as your dashboards.
ChatGPT
your question, in plain language
an answer on governed numbers