Ordergroove MCP: Sample Prompts and Best Practices

Once Ordergroove MCP is connected to your AI client, you can ask about your subscription program in plain language. This article shares sample prompts for common use cases, tips for getting accurate answers, and what to do when the numbers look wrong.

Use cases and sample prompts

Copy any of these prompts into your AI client and swap in your own dates, products, or IDs.

Automate your weekly KPI report

Skip the weekly routine of copying numbers out of dashboards into a spreadsheet. Ask once and get a summary that's ready to share.

• "Pull my Ordergroove KPIs for last week compared with the week before: total subscription revenue (first-time and recurring), new and cancelled subscriptions, active subscribers, and successful and rejected orders. Put it in a table and call out the biggest changes."
• "Give me last month's high-level Ordergroove KPIs as a CSV I can paste into our reporting spreadsheet."

Tip: If your AI client supports scheduled tasks and is connected to tools like Slack or Google Sheets, you can have this report built and delivered automatically each week.

Plan for out-of-stock and discontinued products

Know exactly how many subscribers a product change will affect before it happens, so you can plan inventory, customer outreach, or a product swap.

• "How many active Ordergroove subscriptions include [product], and how many upcoming orders include it over the next 30 days, by week?"
• "Which of my subscription products are out of stock right now, and how many upcoming orders include each one?"
• "Compare my upcoming Ordergroove orders for the next 30 days with current stock levels in [inventory tool], and flag products that could run short." (Requires your AI client to be connected to your inventory tool too.)

Resolve customer support issues

Get to the root cause of a failed order or a missing subscription in one conversation, instead of digging through records or waiting on an escalation.

• "Order [order ID] didn't go through. What happened, and what should we do next?"
• "A customer checked out with a subscription on [date], but no subscription was created. The order ID from checkout is [order ID]. What went wrong?"
• "Walk me through Ordergroove subscription [subscription ID]: product, frequency, next order date, recent orders, and any failed payments."

Find what's driving growth

Go beyond top-line numbers to see which products, frequencies, and subscription types are growing, and where there's room to grow.

• "Which 10 products drove the most recurring subscription revenue in Ordergroove last quarter, and how does that compare with the quarter before?"
• "Break down recurring revenue by subscription frequency for each of the last 6 months. Which frequencies are growing?"
• "Which products bring in the most cross-sell revenue on recurring orders?"
• "Which of my subscription-eligible products have the fewest active subscriptions?"

Monitor cancellations and payment recovery

Spot problems early, product by product, before they show up in your monthly numbers.

• "For each of the last 6 months, compare cancelled subscriptions with active subscriptions for my top 10 products. Which products have the highest share of cancellations?"
• "This month, how many Ordergroove orders were recovered after a failed payment, how many were rejected, and how many are still in retry? Compare with last month."

Tips for accurate answers

Your AI client decides which Ordergroove data to pull and how to interpret it, so the more specific your question, the more accurate the answer.

• Mention Ordergroove. If your AI client is connected to other data sources, like Shopify or a data warehouse, it might pull from one of those instead. Saying "in Ordergroove" points it to the right place.
• Give a date range. "Last month" or "January 1 to March 31" works better than "recently."
• Be specific about what you're counting. Subscribers are customers; subscriptions are what they subscribe to, and one subscriber can have several. Say whether you mean new, cancelled, or active, and whether you want total, first-time, or recurring revenue.
• Use IDs for lookups. Order, subscription, and customer IDs are the fastest way to find a record, since responses don't include customer names or email addresses.
• Ask for metrics, not lists, when you need totals. Lists come back in batches, so counting records from a list can come up short. Focused questions also return less data, which is faster and uses less of your AI plan's usage.
• Ask your AI client to show its work. If a number looks off, ask which data, date range, and filters it used.

When the numbers look wrong

The connector returns your Ordergroove data as-is, so a number that looks wrong usually means your AI client interpreted the question differently than you meant. Common causes:

• It used a different date range, metric, or filter, for example subscriptions instead of subscribers.
• Active subscription and subscriber counts are a snapshot as of the end of the period, so they don't add up across days or weeks.
• Order counts for recent dates can still change while failed orders are being retried.
• It used another connected data source, or counted records from a list instead of asking for a metric.

To check, ask your AI client which data, date range, and filters it used. Then rephrase your question more specifically and compare the result with Analytics in the Ordergroove Admin for the same dates. If the numbers still don't match, contact Ordergroove Support with the question you asked and the answer you got.

Additional reading

• Ordergroove MCP — what Ordergroove MCP is, what data it can access, how to set it up, and how your data stays secure.
• Analytics Agent — ask questions and get insights from your program data; today it runs through this connector.