How to chat with your marketing data in Claude, ChatGPT, Gemini Enterprise and Microsoft Copilot (2026)

To chat with your marketing data, connect your AI assistant to your ad and analytics accounts through Supermetrics. This guide covers Claude, ChatGPT, Gemini Enterprise and Microsoft 365 Copilot, with setup guidance, prompts to try, and accuracy and security checks. The Gemini Enterprise integration is currently available to a limited group of customers; contact Supermetrics to request access.

You ask Claude to sharpen your ad copy, or to pick apart a strategy. But when you want to know which campaigns wasted budget last week, you’re back to logging into separate platforms, pulling reports, and putting it all together manually.

But did you know your AI assistant can help with that question, too? Connect it to your marketing accounts through Supermetrics, and you can analyze campaign performance in plain language, then follow the conversation wherever the findings lead.

Here’s how to get started in Claude, ChatGPT, Gemini Enterprise or Microsoft Copilot, with practical prompts to try and checks to make before acting on an answer. The examples use Supermetrics, a marketing intelligence platform used by 200,000+ companies in 120 countries (as of 2026), with 15 years of experience connecting marketing and sales data.

Key takeaways

• Supermetrics lets you ask plain-language questions about connected marketing data in Claude, ChatGPT, Gemini Enterprise or Microsoft 365 Copilot, without code.
• Supermetrics uses MCP (Model Context Protocol), an open standard for connecting AI assistants to external tools and data, to bring your marketing data into the chat.
• CSV uploads work for one-off analysis, but need refreshing when you want newer data. Supermetrics' AI integrations support 170+ marketing data sources through one connection. The sources and accounts you can use depend on your subscription.
• Supermetrics holds SOC 2 Type II and ISO 27001 certifications, supports SSO (single sign-on) and regional hosting, and never uses customer data to train public models.
• Verdict: By connecting your marketing data to AI via Supermetrics, you can analyze it just by asking natural language prompts.

What does it mean to chat with your marketing data?

Chatting with your marketing data means asking an AI assistant a question in plain language and getting an answer based on data retrieved from your marketing accounts. For example, you might ask:

• "How did my Meta Ads campaigns perform last week?"
• "Which LinkedIn Ads campaigns spent the most without producing a single lead?"

You can then follow up on whatever stands out. Ask the assistant to break down a campaign’s results by ad group, plot the trend or turn the findings into a client update. Each answer gives you something to explore, without starting a new report for every question.

How does an AI assistant access your marketing data?

Supermetrics connects your assistant to your marketing accounts through MCP (Model Context Protocol), an open standard for connecting AI assistants to external tools and data. This lets the assistant retrieve figures from your connected accounts when you ask a question.

Say you ask: “How did my Meta Ads spend change last week compared with the week before?”

Your assistant uses the connection to request spend data for both periods. Supermetrics retrieves those figures from your connected Meta Ads account and returns them to the assistant, which can calculate the difference and explain the results.

You can then ask a follow-up, such as: “Which campaigns drove the increase?” The assistant can request a campaign-level breakdown to investigate further.

Supermetrics provides access to the data; your AI assistant interprets it and shapes the response. That distinction matters: an answer can use the right figures and still draw the wrong conclusion, so check the reasoning before acting.

Supermetrics also manages the MCP server (the service that handles these requests), so you don’t need to build or maintain the connection yourself.

Why not just upload a CSV or connect each platform separately?

Uploading a CSV can work well for a one-off analysis. A connected workflow reduces repeated exports when you need updated data, while one connector makes multiple sources simpler to manage.

The main differences are how you get updated figures, compare channels and manage access. Here’s how the three approaches compare.

How do you connect your marketing data to Claude, ChatGPT, Gemini Enterprise or Microsoft Copilot?

Before you start, check that your Supermetrics subscription includes the AI Chats destination, or begin by starting a free 14-day trial. Then, connect marketing sources from the chat or use accounts already connected in the Hub.

How do you set up Supermetrics for Claude?

1. Open the Supermetrics for Claude page and click "Try in Claude".
2. Sign in to Supermetrics and select the team with your AI Chats subscription or trial.
3. Ask Claude to connect a source such as Google Ads, Meta Ads, GA4, HubSpot or Shopify, then follow the login link to authorize access. You can also use accounts already connected in Supermetrics.

How do you set up Supermetrics for ChatGPT?

1. Open the Supermetrics for ChatGPT page and click "Try for free".
2. Sign in to Supermetrics with the account you use for the Hub, and select the team with your AI Chats subscription or trial.
3. Connect a marketing source when prompted, or select an account already connected in Supermetrics.

How do you connect Gemini Enterprise?

The Supermetrics Marketing Intelligence Agent for Google Gemini Enterprise is available to a limited group of customers. Access requires enablement through the Supermetrics team.

1. Contact Supermetrics to request access and discuss your organization's setup.
2. Once enabled, follow the team's onboarding instructions to sign in and connect or select the marketing accounts you want to analyze.

How do you set up Supermetrics in Microsoft 365 Copilot?

Use the Supermetrics Marketing Analyst agent in Microsoft 365 Copilot or Microsoft Teams. Your organization may need to approve the agent before you can add it.

1. In Copilot, open "All agents", find "Supermetrics Marketing Analyst" and click "Add". Alternatively, open the Supermetrics for Copilot product page and click "Try in Copilot".
2. If you can't add the agent, ask your Microsoft 365 administrator to approve or deploy it.
3. Follow the agent’s onboarding instructions to sign in to Supermetrics and connect or select your marketing accounts. If you don’t have an active Supermetrics subscription, a free 14-day trial starts automatically when you first connect a data source.

Check your first answer

Once you’re set up, it’s time to check everything is working as it should. Select one account and ask: "What was the total spend in my [account name] Google Ads account last week? Show the account, dates and currency used."

Compare the answer with Google Ads before adding more sources. In Copilot, address the prompt to Supermetrics Marketing Analyst.

What questions should you ask your marketing data?

Ask your marketing data the questions you'd normally answer with exports and spreadsheets: quick performance checks, cross-channel comparisons, anomaly and pacing checks, and budget questions.

Start with a question you already need to answer, whether that’s checking last week’s results or preparing for a budget meeting. Try one of the prompts below, replacing the platforms and targets with your own.

Quick performance checks

• "Compare Meta Ads spend, purchases and cost per purchase last week with the week before. Highlight the biggest changes."
• "Show yesterday’s spend, reported conversions and cost per conversion for Google Ads and Meta Ads separately. State which conversion metric you used for each."

Cross-channel comparisons

• "Compare ROAS (return on ad spend) across Google Ads and Meta Ads for the last 30 days and flag campaigns below 2x. Use each platform’s reported ROAS and flag any differences in attribution settings."
• "Compare Google Ads and LinkedIn Ads spend and reported leads last quarter. Calculate cost per lead for each and show your workings."

Anomaly and pacing checks

• "Which Google Ads campaigns had the biggest percentage increase in CPC (cost per click) last week? Show their spend, clicks and impressions for both weeks."
• "Our planned Google Ads spend this month is £20,000. Based on month-to-date spend, are we on track to exceed it? Show your projection."

Budget and forecasting questions

• "Using the last 90 days of Meta Ads spend, estimate our spend for the rest of this month. Explain your method and assumptions."
• "Model moving £1,000 from LinkedIn Ads to Google Search using last quarter’s average cost per lead. Show the estimated change in leads and explain the limitations."

How do you write prompts that get sharper answers?

For sharper answers, specify the account, metrics and dates you want to analyze. Add relevant business context, ask follow-up questions and check how the assistant calculated its results.

A useful prompt gives your assistant enough direction to investigate the right thing. You don’t need a long brief: a few specific details can make the difference between a broad summary and an answer you can use.

1. Be specific about the data. Name the platform and account, the metrics you want, and the date range. Include a comparison period if you want to understand a change. “Compare cost per purchase in our UK Meta Ads account last week with the week before” gives the assistant more to work with than “How are ads doing?”
2. Explain what matters to your business. Say what counts as a valuable result. If you care about qualified leads, make that clear rather than letting the assistant treat every form submission as equally useful. Mention recent changes, such as a promotion or tracking outage, that could affect its interpretation.
3. Follow up on what stands out. You don’t need to fit the whole investigation into one prompt. Start with a performance check, then ask for a breakdown by campaign or ad group. If one campaign’s costs rose sharply, ask to see its underlying figures for both periods.
4. Ask for the workings. Request the fields, inputs and formula behind calculated metrics. For forecasts or budget scenarios, ask the assistant to explain its assumptions and what the calculation cannot tell you.
5. Say how you’ll use the answer. A client update needs a different response from a detailed campaign review. Ask for the format you need, such as a short explanation for a client or a table of campaigns to investigate.

You shouldn’t have to explain the same reporting rules in every new chat. That’s why Supermetrics’ Business Context feature (released in September 2026) saves reusable guidance for your team, such as conversion definitions, reporting preferences and which accounts belong to each client.

For example, tell your assistant: “For this account, our lead-quality metric is qualified leads, not all form submissions. Remember that.” The assistant can offer to save this guidance for future conversations.

Team owners, admins and editors can save or update Business Context. Saved guidance follows the same account-access rules as your data.

What can you do beyond asking questions?

You can turn your analysis into reports and presentations, build dashboards that stay updated in Supermetrics Studio, and use supported AI integrations to create or manage ad campaigns.

Once you’ve checked the findings, put them to work. The next step might be explaining a result to your client, creating a view your team can return to or making a campaign change.

Turn findings into reports and presentations

Ask your assistant to turn the analysis into a weekly performance summary, a client update or a presentation outline. Tell it who will read the output and what they need to know.

For example: “Turn these findings into a short client update. Explain the biggest performance change, show the supporting figures and suggest what we should investigate next.” Review the draft before sharing it.

Layer used Supermetrics for Claude to turn a 10-hour report into 20-minute tasks, showing the value of AI-powered marketing reporting for agencies.

Build dashboards your team can keep using

A useful analysis doesn’t have to stay buried in a chat. In Claude or ChatGPT, you can build a dashboard from your connected data and ask the assistant to push it to Supermetrics Studio, where it can refresh and be shared with your team.

Your AI Chats subscription lets you push dashboards to Studio, then view and share them. To edit them in Studio and access its full feature set, you need a Supermetrics Studio destination subscription. See the Supermetrics Studio guide for the steps.

Create and manage ad campaigns

With campaign management enabled in a supported integration, you can move from investigating performance to making changes. For example, ask your assistant to draft a new campaign, pause a campaign or adjust its budget.

You must grant write access to the relevant accounts and choose when human approval is required. New campaigns are created paused (unless you specifically turn off this feature), and Campaign History records changes and lets you reverse supported actions.

Is it accurate and secure to chat with your marketing data?

Connecting your marketing accounts through Supermetrics gives AI answers a basis in your actual performance data and lets you control access. You still need to check important findings and review your AI provider’s privacy settings.

How does connected marketing data improve accuracy?

An assistant can’t know what you spent on Meta Ads last week unless it has the figures. Supermetrics retrieves data from your connected accounts, giving the assistant real numbers to analyze rather than leaving it to guess. Clear prompts and relevant business context help it use those numbers appropriately.

That reduces guesswork, but it doesn’t eliminate mistakes. An assistant can still misinterpret a metric, overlook a data gap or draw an unsupported conclusion. Start with a small query you can check against the source platform, then expand once you’re confident in the results. Our guide to reducing AI hallucinations explains this approach, with lessons from marketing agency Layer.

Is it safe to connect your marketing data to AI?

With Supermetrics’ AI chats, the AI tool requests data through Supermetrics rather than logging into your ad platforms directly. You authorize the source connections, and you can revoke access. Supermetrics encrypts data in transit, while Enterprise single sign-on gives admins central control over access.

Your AI provider also handles the data returned to the chat, so its policies matter too. Training and retention rules depend on the product, plan and settings you use; check that these meet your company’s requirements before connecting accounts.

Ready to chat with your marketing data in AI?

Connect one account to the assistant your team already uses, then try the test question above. Once you've checked the result, ask a follow-up about the campaigns or trends you want to investigate.

Start a 14-day trial of Supermetrics for Claude or Supermetrics for ChatGPT to begin.

Frequently asked questions