How to Connect Criteo to ChatGPT (2 Steps, No Code)

The fastest way to connect Criteo to ChatGPT is Windsor.ai’s native app: link your Criteo account to Windsor, then connect the same app inside ChatGPT. Once connected, ChatGPT can analyze your Criteo retargeting performance, impressions, clicks, spend, revenue, and ROAS, in plain language, without you exporting a single report.

Criteo’s own dashboard shows solid campaign and product-level retargeting data, but it lives on its own. Comparing that spend against your Google Ads or Meta Ads budget still means exporting a CSV from each platform and rebuilding the same pivot table every week, and the moment you export it, the numbers are already out of date.

2 steps to connect Criteo to ChatGPT

Takes under a minute, no code or manual CSV work required.

Follow our quick step-by-step guide to integrate your data into ChatGPT through Windsor MCP.

Prerequisites

To use Windsor MCP with ChatGPT, make sure you have the following:

• An active Windsor.ai account (free or paid)
• A Criteo account with at least one active campaign
• An active ChatGPT account

Step 1. Connect Criteo to Windsor.ai

• Create your Windsor.ai account at onboard.windsor.ai and start for free.
• Select Criteo as your data source and sign in with your Criteo credentials to authorize access.
• Connect more than one Criteo account if you manage several advertisers or markets.

Step 2. Connect the Windsor.ai app in ChatGPT

• Open the Windsor.ai app page in ChatGPT and click Connect.
• Approve access. Windsor MCP is now connected.

List the data sources connected to my Windsor account.
Do you see Criteo there?

Once ChatGPT recognizes your Criteo data, you’re ready to query it in natural language.

The metrics that matter for Criteo

Metric	What it tells you
Spend	How much budget is going into your Criteo campaigns, the baseline for judging whether the results are worth the cost.
Impressions	How many times your retargeting ads were shown to shoppers, a read on how much visibility your campaigns are getting.
Clicks	How many people clicked through to your site after seeing an ad, an early signal that your creative and targeting are working.
Click-Through Rate	The share of ad displays that turned into clicks, useful for comparing creative and audience performance side by side.
Frequency	How many times, on average, each shopper saw your ad, a high number can be an early warning sign of ad fatigue.
Omnichannel ROAS	Revenue earned for every dollar spent across devices, the clearest single number for judging campaign profitability.
Revenue	Total sales revenue attributed to Criteo within 30 days of a click, the direct dollar payoff of your retargeting spend.
Transactions	How many purchases your Criteo ads actually drove, turning clicks into a concrete sales count.
Average Cart	The average order value from purchases your ads influenced, useful for understanding the type of customer retargeting brings back.
Cost Per Order	How much you’re paying, on average, to generate one sale, a key efficiency check on whether retargeting spend is scaling sensibly.

What to ask ChatGPT: prompt ideas

Campaign performance

Show my total Criteo spend and revenue by campaign for the last 30 days.

Break down my Criteo spend and Omnichannel ROAS by campaign for the last 90 days and flag any campaign where ROAS is below 2.

Post-view vs. post-click impact

Compare post-view revenue and post-click revenue by campaign for the last 30 days and tell me which campaigns rely most on post-view revenue.

Advertiser and conversion analysis

Summarize my Criteo transactions and average cart value by advertiser for this month.

Rank my Criteo campaigns by click-through rate and conversion rate for the last 14 days.

Windsor’s connection to Criteo is read-only, so ChatGPT can analyze your retargeting performance but can’t create, pause, or edit campaigns in Criteo.

Bonus: analyze Criteo alongside your full stack

Windsor connects 350+ sources, so Criteo doesn’t have to live in its own silo. A few realistic blends:

• Blended ROAS across paid channels. Combine Criteo spend and revenue with Google Ads and Meta Ads to see one cross-channel ROAS number instead of judging retargeting spend in isolation.
• Retargeting versus real sales. Pair Criteo’s post-click revenue with GA4 or your e-commerce platform’s order data to see how much of that revenue lines up with actual completed purchases, not just Criteo’s own tracking.
• Stock-aware retargeting. Blend Criteo campaign data with your product or inventory feed to spot when you’re spending on ads for items that are low on stock or already sold out.

Other ways to connect Criteo to ChatGPT

Windsor is the fastest no-code way to do this, but two other methods exist if you’d rather not use a connector.

Manual CSV export

Export a CSV from the Criteo dashboard and upload it to ChatGPT for a one-off analysis. It works for a single question, but it’s a static snapshot: the moment you download it, it’s already out of date, and you’re repeating the export every time you want a fresh number.

Zapier or Make

Build an automation that exports Criteo data on a schedule to a file ChatGPT can read. It removes the manual export step, but you’re now maintaining a workflow, and it’s still a delayed copy rather than a live connection.

Both avoid a direct connector, but neither gives ChatGPT a live view of your Criteo account the way Windsor’s connector does.

Conclusion

Setup takes a few minutes: connect Criteo to Windsor, add Windsor MCP to ChatGPT, and start asking questions about your retargeting performance directly. No CSV exports, no stale numbers, no rebuilding the same pivot table every week.

🚀 Start for free and get your Criteo data live in ChatGPT: onboard.windsor.ai/app/criteo