Microsoft Copilot
Answers, not guesses

Ask Copilot about your store.
Answered inside Microsoft 365.

Copilot doesn't have your commerce data, so it guesses. Polar fixes that. Connect once and ask about your live Shopify, Meta, Google, Microsoft / Bing Ads, TikTok, Klaviyo and Amazon data in plain English, right inside Excel, Teams, Power BI and PowerPoint. Every answer runs through Polar's governed semantic layer, so it matches your bank account, not a guess.

/
What It Is

One governed data layer, exposed to Copilot

Polar ingests and models 50+ sources into one metric layer, then exposes it to Copilot over standard MCP, with OAuth. So Copilot answers real business questions instead of guessing.
Connects to your whole stack
Shopify Amazon Meta Ads Google Ads Microsoft / Bing Ads GA4 Klaviyo Recharge TikTok +50 sources, one metric layer

Unified data

One query spans Shopify, paid media, email and analytics, deduped, attribution-ready and consistent across every brand view.

Core tools

A focused toolset Copilot composes into real reports.

get_context generate_report get_metrics list_dimensions get_dimension_values list_dashboards get_dashboards create_custom_metric create_custom_dimension

Governed & accurate

Answers come from Polar's modeled metrics, the same numbers as the Polar app, not hallucinated SQL. The data leadership already trusts.

Multi-brand

Run several stores or manage clients as an agency? Switch across every brand view in a single connection, built for portfolio work from day one.

Setup Guide

Connect Copilot to your data, step by step.

Connect your data in Polar, add Polar as a connector in Copilot, then just ask. A few minutes, no code, no API keys to manage.
Polar Analytics

Polar Analytics account

Connect Shopify, Meta, Google, Klaviyo and more. The Polar MCP exposes clean, attributed live data, the source of truth Copilot reads from.

Connect your data →
Microsoft Copilot

Microsoft Copilot + Copilot Studio

Add Polar as an MCP tool in Copilot Studio, build an agent once, then use it across Microsoft 365: Excel, Teams, Outlook, PowerPoint, Power BI and Copilot chat.

Get the Polar MCP endpoint →
Polar multi-brand

Multiple stores? One connection

Whether you run several stores yourself or manage clients as an agency, switch across every brand view inside a single Polar connection. Build the agent once, then get governed answers for any store, no separate setup each time.

Talk to us about multi-brand →
1

Connect your data in Polar ~5 minutes

Build the semantic layer Copilot will read from.
How to
1
In Polar, connect Shopify, your ad platforms, Klaviyo, Amazon and GA4. Polar's semantic layer turns raw exports into governed metrics: blended ROAS, true CAC, contribution margin, defined once and consistent everywhere.
2
Make sure your data sources are connected, then wait for the first sync. The initial historical sync can take up to 24 hours for a new Polar account.
Why it matters: the answer is only as good as the data behind it. Polar's layer means the numbers Copilot returns in Excel or Teams match your bank account, not a cookie estimate or an LLM's guess.
Connect your data in Polar
2

Build your Polar agent in Copilot Studio ~5 minutes

Add Polar as an MCP tool, then publish. No custom build.
How to
1
In Copilot Studio, create an agent and add Polar Analytics as an MCP tool, using Polar's one-URL MCP endpoint with OAuth.
2
Give the agent its role and knowledge (Polar's metric layer, 50+ sources), then publish it to Microsoft Teams and Microsoft 365 Copilot.
3
Use the agent wherever you work: ask in Excel, Teams, Outlook or Copilot chat. Build it once, reuse it everywhere.
Native MCP fit: Copilot Studio speaks MCP, so Polar's existing endpoint connects as an agent tool with no custom build. The same one-URL endpoint already powers Polar in Claude's connector directory.
Get the Polar MCP endpoint
3

Ask anything, inside Microsoft 365 the payoff

Ask Copilot in Excel, Teams or chat. Polar powers the answer.
What this unlocks

A real answer, not a hallucination.

Ask Copilot like you always do, in Excel, Teams or chat. Polar's connector pulls the live, attributed numbers from your semantic layer and Copilot reasons over them: blended ROAS, true CAC, contribution margin, LTV cohorts, all pre-computed, so a number in your workbook is one you can trust.

"Polar's semantic layer is way cleaner structure, way fewer hallucinations. The most delightful MCP experience." How RipnDip uses Polar's MCP vs the alternatives.

A few of the questions you can ask:

ASKBuild an Excel table of true CAC by channel last 90 days.
ASKWhy did blended ROAS drop last week?
ASKDraft a Teams update with yesterday's revenue and margin.
ASKWhich SKUs will stock out before their lead time?
ASKBuild a Power BI dataset of MER by month.
ASKDraft a board slide of this month's KPIs vs plan.
See the prompt library
Get started

Only Polar gives Copilot clean, attributed data to answer from.

Connect your data in Polar, add the Polar MCP connector to Copilot, then just ask, inside Excel, Teams and Power BI. A real answer in seconds, no code.

See It In Action

Build the agent once. Ask everywhere.

Build a Polar agent in Copilot Studio, then ask in Excel, Teams, Outlook or Copilot chat. Governed Polar metrics, in the surface you already work in. No export, no copy-paste, no stale screenshots.
Copilot Studio Publish

Weekly Growth Exec Review

Custom agent · governed by Polar Analytics
Instructions
Role — You are the Head of Growth. Produce a sharp weekly executive review using Polar Analytics as the sole analytics source. Lead with what happened, then why, then what to do.
Tools
Polar Analytics · MCP + Add tool
Knowledge
Polar metric layer · 50+ sources
Channels
Microsoft Teams Microsoft 365 Copilot + Add channel
What the agent returns · the weekly review
Weekly Growth Exec Review May 25–31 · vs prior week
1 What happened

More traffic and orders, but less revenue and thinner efficiency. Net sales fell to $1.32M (-10.6%) even as orders rose +2.2% and sessions +5.4%. AOV slipped -2.6%, gross margin compressed 62.7% → 59.0%, and blended ROAS dropped 14.20x → 12.63x.

Not a demand problem, a quality-of-demand problem: more visits and orders, but lower value per order, thinner margin and weaker media efficiency.
2 Why it happened
DriverSignalRead
TrafficSessions +5.4%Demand is healthy
ConversionCVR -1.9%, AOV -2.6%Order quality softening
MediaROAS -11.1%Spend outran return
3 What to do
  • Pause bottom-quartile Meta & Google ad sets before the weekend, spend grew faster than return.
  • Push 'box of 12 / 20' bundle variants on PDPs to pull AOV back up.
  • Audit margin by channel, protect contribution, not just top-line revenue.
In Excel · ask Copilot, get a live model
Excel Pull last week's KPIs from Polar and compare to the prior week
 ABCD
1MetricLast weekPrev weekWoW
2Net sales$1.32M$1.47M-10.6%
3Orders12,72212,453+2.2%
4AOV$126.11$129.45-2.6%
5Sessions657,352623,793+5.4%
6Conversion rate1.72%1.75%-1.9%
7Ad spend$118.5K$117.4K+0.9%
8Blended ROAS12.63x14.20x-11.1%
Polar · WeeklyLive cells, Copilot re-queries Polar's metric layer on refresh, so models built on top stay current.
Three Pieces, One Flow

You ask. Polar answers. Copilot explains.

Copilot is the interface you already use. Polar's semantic layer turns your question into a governed, accurate query, using your definitions of ROAS, CAC and margin, on clean attributed data.
01 · Ask

Copilot

Inside Microsoft 365, where you already work

02 · Translate

Polar MCP

Your governed metrics

03 · Answer

Your live data

Clean, attributed, in Excel, Teams, PowerPoint and more

Your Prompt Library

61 prompts Copilot can run on your live data

Filter by what you need or who it is for, copy the prompt, paste it into Microsoft Copilot with Polar connected. Every answer comes from your governed, attributed data, right inside Excel, Teams, Outlook, PowerPoint and Power BI.
Type
Audience
61 of 61
The Foundation

Copilot didn't have your data. Now it does.

Out of the box, Copilot guesses, or you paste in a CSV and hope it reads it right. Polar's semantic layer feeds it your live, attributed numbers directly, so every answer is finance-grade: matches your bank account, never a hallucination. No SQL, no exports, no data team.
Connect Polar to Copilot
Get started · connect once · ask anything

Copilot didn't have your numbers. Now it does.

Connect Polar to Copilot, then just ask, right inside Excel, Teams and Power BI. Every answer runs through Polar's governed semantic layer: finance-grade, no SQL, no exports.

Phase 1 ~5 minConnect your data in PolarConnect Shopify and your stack. Polar's semantic layer turns it into governed metrics, then wait for the first sync.
Phase 2 ~5 minConnect Polar to CopilotAdd Polar's MCP connector in Microsoft 365, click Connect and sign in to Polar. No setup.
Phase 3 the payoffAsk anything, inside Microsoft 365Type your question in Excel, Teams or Copilot chat. Polar pulls the live, attributed numbers and Copilot answers.