Most ecommerce marketers use ChatGPT to write subject lines, generate email copy, or come up with campaign ideas. But writing an email is only one part of email marketing.
The harder decisions often come before that. Which customers should receive a campaign? Which products should you promote? Does the audience need a discount? And how do you know whether the campaign actually worked?
These are decisions you can make using data that’s already available in your ecommerce store and email marketing platform. With access to the right data, ChatGPT can help identify these patterns, recommend campaigns, build customer segments, choose products and offers, and analyze results. It can also help with the more familiar tasks of writing emails and improving automated flows.
Here are 23 ChatGPT prompts for ecommerce email marketing, covering everything from identifying campaign opportunities to planning, creating, and evaluating campaigns.
Before you use these prompts, make sure ChatGPT can access the right data
The quality of ChatGPT’s recommendations depends on the information you provide.
For example, if you ask ChatGPT to identify customers who are likely to reorder soon, it needs their previous purchase dates and order history. If you ask whether a customer segment should receive a discount, it needs information about previous discount usage, purchase patterns, and product margins.
Without that information, ChatGPT can suggest general strategies, but it cannot reliably tell you what makes sense for your customers.
1. Connect your ecommerce store
If you’re using Shopify, you can provide ChatGPT with access to store data through a supported MCP integration or an API-based workflow.
The most relevant information includes:
- Customer data: Customer profiles, lifecycle stage, and purchase history
- Order data: Previous purchases, order values, order dates, and discounts used
- Product data: Product names, categories, prices, and related products
- Inventory data: Stock availability, inventory levels, and product sales
This information helps ChatGPT identify opportunities based on what customers have actually purchased and what the business has available to sell.
2. Connect your email marketing platform
Your store data explains what customers purchase. Your email service provider (ESP) provides additional information about how they interact with your email campaigns.
You can connect a supported ESP through an MCP or API workflow to make the following information available:
- Campaign history: Emails sent to different audiences and when they were sent
- Email engagement: Opens, clicks, and other available engagement events
- Campaign performance: Revenue, orders, and conversions attributed to campaigns
- Automated flows: Existing welcome, abandonment, replenishment, and win-back sequences
For example, a customer may have purchased three times during previous promotional campaigns but never at full price. Combining Shopify order history with ESP campaign data can help you understand whether that customer responds primarily to discounts.
3. You can also upload the data manually
You don’t need a live integration to use these prompts.
You can export the relevant customer, order, product, or campaign data into a CSV file and upload it to ChatGPT. For smaller datasets, you can also paste the information directly into the conversation.
Before uploading, remove unnecessary personal information such as customer names, email addresses, phone numbers, and other identifying details. Use anonymized customer IDs where individual records need to be compared.
You also don’t need to upload your entire store database for every prompt. A customer segmentation prompt may only require purchase history and discount usage, while an offer-selection prompt also needs product margins and previous campaign results.
One important rule: If the necessary information isn’t available, ChatGPT should identify what’s missing rather than make up a recommendation. Keep that distinction in mind when reviewing its output.
Find campaign opportunities
Before planning another email, it’s worth understanding where the strongest opportunities actually exist. Customer purchases, product performance, and engagement data can help identify campaigns worth running.
Prompt 1: Find campaign opportunities from customer purchase data
Your order history can reveal customers who haven’t returned for a second purchase, repeat buyers who may be approaching their usual reorder date, and customers who respond differently to discounts.
I’m going to give you customer and order data from my ecommerce store. Analyze it and identify the strongest email campaign opportunities.
Look for useful patterns such as:
— first-time buyers who have not purchased again
— customers approaching their normal reorder window
— customers whose purchase frequency is declining
— high-value customers worth targeting separately
— category buyers who may be relevant for cross-sell
— customers who usually buy on promotion versus at full price
For each opportunity, return:
— the customer group
— the behavior you identified
— why it creates a campaign opportunity
— the recommended campaign
— the products or categories to feature
— the best primary metric to track
Do not suggest a campaign unless the data gives you a clear reason.
Customer and order data: [PASTE DATA]
Prompt 2: Find campaign ideas from product and inventory data
Customer demand isn’t the only reason to run a campaign. Use sales trends, inventory levels, and product margins to identify products that could benefit from email promotion.
Analyze the following product, sales, and inventory data and identify where an email campaign could help the business.
Look for situations such as:
— healthy inventory but slowing sales
— products with strong margins and room for more demand
— new products that need awareness
— products with strong repeat-purchase potential
— categories showing unusual growth or decline
— bundles or related products that could be promoted together
For every recommendation, explain:
— what changed
— why email is an appropriate response
— which customers should receive the campaign
— which products should be featured
— whether an offer appears necessary
Do not recommend pushing products simply because inventory is available.
Data: [PASTE DATA]
Prompt 3: Find customers approaching their normal reorder window
One customer might reorder every 30 days, while another takes 60 days or longer. Use this prompt to compare previous purchase intervals with the time since the latest order and identify potential reorder opportunities.
I’m going to give you customer order history. Identify customers who appear to be approaching their normal reorder window.
Compare the time since the most recent purchase with each customer’s previous purchase intervals.
Do not assume all customers follow the same reorder cycle.
For each customer or segment, return:
— usual reorder interval
— days since last purchase
— product usually reordered
— whether the customer appears early, on time, or late relative to their normal cycle
— recommended email timing
— recommended email angle
Order history: [PASTE DATA]
Prompt 4: Find first-time buyers who are ready for a second-purchase campaign
Identify first-time buyers worth targeting based on what they purchased, when they purchased, and what they’ve shown interest in since.
Analyze these first-time buyers and identify which groups are the strongest candidates for a second-purchase email campaign.
Consider:
— what they bought
— how long ago they purchased
— normal second-purchase timing for similar customers
— categories or products they have browsed since purchasing
— whether they have purchased with or without discounts
— relevant complementary or replenishment products
Group customers into sensible audiences and recommend the next campaign for each group.
Do not assume every first-time buyer should receive the same follow-up.
Data: [PASTE DATA]
Prompt 5: Find customers beginning to become inactive
A customer doesn’t have to stop purchasing completely to show signs of declining interest. Spot customers whose purchase frequency or engagement is declining so you can reach them before they become inactive.
Analyze this customer purchase and engagement data and identify customers whose behavior appears to be declining relative to their own normal pattern.
Look for:
— missed expected repurchase windows
— declining purchase frequency
— reduced browsing or email engagement
— customers still browsing but no longer purchasing
— meaningful changes from previous category activity
For each group, recommend whether the brand should intervene now or wait.
If intervention is justified, suggest the most relevant re-engagement angle.
Data: [PASTE DATA]
Choose the audience, products, and offer
Once you’ve identified a campaign opportunity, the next step is deciding who should receive it, what products to feature, and whether an incentive is needed.
Prompt 6: Build the right audience for a campaign
A new product or promotion may be relevant to only part of your customer base. Identify which customers should receive a campaign, who should be excluded, and whether different groups need different messages.
I’m planning an ecommerce email campaign with the following goal:
[CAMPAIGN GOAL]
Products or category:
[PRODUCTS]
Using the customer data below, recommend:
— who should receive the campaign
— who should be excluded
— whether different customer groups need different versions
— the customer signals that justify each decision
Consider purchase history, category interest, lifecycle stage, recency, frequency, discount behavior, and recent campaign exposure.
Avoid creating segments that are different on paper but would receive essentially the same message.
Customer data: [PASTE DATA]
Prompt 7: Find useful customer segments from purchase behavior
Two customers with the same lifetime spend may still shop very differently. Group customers by purchasing habits and identify which segments would benefit from different email strategies.
Analyze this customer purchase data and identify customer segments that are meaningfully different enough to deserve different email treatment.
Consider:
— purchase frequency
— average order value
— category affinity
— full-price versus promotional purchasing
— time since last purchase
— repeat-purchase behavior
— product combinations
For each segment, explain:
— what makes the group distinct
— how their email strategy should differ
— which campaign types are most relevant
Do not create segments unless the difference would change the marketing action.
Data: [PASTE DATA]
Prompt 8: Decide whether a campaign needs a discount
Offering a discount can increase conversion, but it can also reduce margins on orders customers might have placed at full price. Use previous purchases, discount behavior, and product margins to decide whether an incentive is necessary.
I’m planning an email campaign for:
Audience: [AUDIENCE]
Product/category: [PRODUCT]
Campaign goal: [GOAL]
I’ll also give you customer purchase behavior, previous campaign results, discount history, product margin, and inventory.
Recommend whether this campaign should use:
— no incentive
— percentage discount
— fixed-value discount
— free shipping
— bundle
— threshold offer
Do not recommend a discount simply because it might increase conversion.
Consider:
— whether these customers usually buy at full price
— their previous response to promotions
— available margin
— purchase intent
— whether a weaker incentive could achieve the same goal
Explain your recommendation and identify any customer groups that should receive a different offer.
Data: [PASTE DATA]
Prompt 9: Choose the best offer for a customer group
Once you’ve decided that a campaign needs an incentive, you still have to choose the right one. Compare discounts, bundles, and other incentives to find an offer that balances conversion potential with profitability.
Compare the following offer options for this ecommerce audience:
[LIST OFFERS]
Audience: [AUDIENCE]
Product/category: [PRODUCT]
Goal: [GOAL]
Using the customer behavior and product economics below, rank the offers from strongest to weakest.
Evaluate each based on:
— likely impact on conversion
— margin impact
— average order value
— customer discount sensitivity
— whether customers might purchase without an incentive
— risk of training customers to wait for promotions
Recommend the smallest incentive likely to change behavior.
Data: [PASTE DATA]
Prompt 10: Choose which products to promote
A bestselling product isn’t necessarily the best recommendation for every audience. Select products based on customers’ purchase history and interests rather than simply promoting bestsellers.
I’m planning a campaign for this customer group:
[AUDIENCE]
Using their purchase history, browsing behavior, category interest, and the product data below, recommend which products should be featured.
Rank the products and explain:
— why each product is relevant
— whether it should be the hero product or a secondary recommendation
— which customers should not see it
— whether there are better cross-sell or replenishment alternatives
Avoid recommending products solely because they are best sellers.
Data: [PASTE DATA]
Plan the campaign
Once you’ve decided which customers to target and what to promote, you need to turn those decisions into a campaign that can actually be executed.
Prompt 11: Turn a campaign goal into a full campaign brief
A campaign brief should give your marketer and designer enough information to create the email without having to decide the strategy themselves. Turn your campaign objective, audience, products, and offer into a brief that they can build on.
Build a complete ecommerce email campaign brief using the information below.
Goal: [GOAL]
Audience: [AUDIENCE]
Product/category: [PRODUCT]
Offer: [OFFER]
Brand context: [BRAND]
The brief should include:
— campaign objective
— audience
— customer insight
— product selection
— offer
— main message
— supporting proof
— CTA
— recommended timing
— anything that should be avoided
Keep the brief specific enough that a copywriter or designer could create the campaign without guessing what the strategy is.
Prompt 12: Choose the best campaign angle
The same product can be promoted through different marketing angles. Explore different ways to position a campaign and identify which angle is most relevant to your audience.
I’m planning a campaign for:
Product: [PRODUCT]
Audience: [AUDIENCE]
Goal: [GOAL]
Generate 5 distinct campaign angles.
Consider approaches such as:
— product benefit
— customer problem
— education
— newness
— popularity
— social proof
— founder story
— urgency, only if there is a real deadline
For each angle, explain:
— why it could work for this audience
— the core message
— the best evidence or proof to include
— any risk or weakness in the angle
Then recommend the strongest option.
Prompt 13: Create different versions for different customer groups
Not every customer segment needs the same introduction to a product. Adapt the message, products, or offer for different customer segments while keeping the campaign’s overall objective the same.
I have one campaign objective:
[GOAL]
But I’m targeting these customer groups:
[SEGMENTS]
Create a different campaign approach for each group.
The overall campaign should remain recognizable as the same campaign, but adapt where useful:
— message
— product selection
— proof
— offer
— CTA
— level of education
Only change something when the customer context gives you a reason.
Prompt 14: Decide when a campaign should be sent
The right time to send an email depends on more than the day of the week. Determine the right sending window based on customer activity, purchase patterns, and recent email exposure.
Recommend the best timing for this ecommerce email campaign.
Campaign: [CAMPAIGN]
Audience: [AUDIENCE]
Consider:
— lifecycle stage
— time since last purchase
— recent browsing or cart activity
— expected repurchase timing
— recent email exposure
— campaign deadline
— any fixed business event
Tell me:
— recommended send window
— whether all customers should receive it at the same time
— which customers should receive it earlier or later
— who should be suppressed
Customer behavior: [PASTE DATA]
Prompt 15: Review an ecommerce campaign calendar
It’s easy to fill a campaign calendar with product announcements, promotions, and reminders without checking whether the same customers are receiving too many emails. Review your campaign calendar to identify repetitive emails, overlapping audiences, and missed campaign opportunities.
Review the following ecommerce email campaign calendar.
Identify:
— repetitive campaigns
— too many promotional sends
— missing lifecycle opportunities
— audiences that may be receiving too much email
— gaps where customer or business data suggests a campaign opportunity
— campaigns that could be narrowed or personalized
For every change you recommend, explain why.
Do not add campaigns simply to fill empty dates.
Calendar: [PASTE CALENDAR]
Customer/business context: [PASTE CONTEXT]
Write and create the email
Once the strategy and campaign brief are ready, ChatGPT can help turn them into the actual email. The key is to give it enough context about the product, audience, offer, and brand voice.
Prompt 16: Write a complete ecommerce email from a campaign brief
Instead of asking ChatGPT to write a generic promotional email, give it a completed campaign brief. Use this to generate email copy that follows your brand voice and campaign objective.
Write the ecommerce email described in the campaign brief below.
Follow these rules:
— use the supplied brand voice
— keep the message focused on one primary job
— use plain, natural language
— do not invent product claims
— do not use fake urgency
— keep the CTA clear
— do not add a discount unless it is in the brief
Return:
— subject line
— preview text
— headline if needed
— body copy
— CTA
Campaign brief: [PASTE BRIEF]
Brand voice examples: [PASTE EXAMPLES]
Prompt 17: Generate subject lines and preview text
Subject lines should communicate the campaign’s actual message, not just attract attention. Generate different subject-line and preview-text combinations and identify which are most relevant to your audience.
Generate 10 subject-line and preview-text pairs for this ecommerce campaign.
Campaign: [CAMPAIGN]
Audience: [AUDIENCE]
Main angle: [ANGLE]
Create genuinely different approaches rather than changing one or two words.
Include a mix of:
— direct
— product-led
— benefit-led
— curiosity-led
— offer-led, if applicable
Avoid clickbait, fake urgency, and vague subject lines.
Rank the top 3 and explain why they fit this audience.
Prompt 18: Write a product launch email
A product launch email should explain what’s new and why the target audience should care.
Write a product launch email using the information below.
Product: [PRODUCT]
Audience: [AUDIENCE]
What is new/different: [DIFFERENTIATION]
Main customer benefit: [BENEFIT]
Proof: [PROOF]
Offer, if any: [OFFER]
The email should help the customer quickly understand:
— what launched
— why it matters
— why it may be relevant to them
— what they should do next
Do not overload the email with every product feature.
Improve flows and lifecycle emails
Automated email flows are useful because they respond to specific customer actions or stages in the buying journey. But they still need to be reviewed for timing, relevance, and unnecessary messages.
Prompt 19: Review a welcome email series
A welcome series should help new subscribers understand the brand and move toward their first purchase.
Review this ecommerce welcome series.
The primary goal is to move a new subscriber toward their first purchase.
For each email, identify:
— its job
— whether it adds something new
— whether the timing makes sense
— whether the signup promise is being fulfilled
— whether the customer is getting enough information to make a purchase
— whether any email is repetitive or unnecessary
Also recommend:
— behavioral branches
— suppression rules
— the correct exit condition
Welcome series: [PASTE SERIES]
Prompt 20: Improve a cart-abandonment sequence
Cart abandonment emails shouldn’t automatically assume the customer left because of the price. Review your cart recovery emails to improve their timing, messaging, discount strategy, and exit conditions.
Review this ecommerce cart-abandonment sequence.
Evaluate:
— timing
— number of emails
— how each message progresses
— whether customer history should change the message
— whether discounts are being introduced too early
— whether the sequence stops after purchase
Do not assume you know why the customer abandoned.
Instead, use known customer history and current behavior to recommend the most relevant recovery angle.
Sequence: [PASTE SEQUENCE]
Customer data: [PASTE DATA]
Prompt 21: Build or improve a replenishment sequence
Replenishment emails are most useful when they arrive close to the time a customer is likely to need another purchase. Plan replenishment reminders around product usage and each customer’s previous reorder patterns.
Build or review a replenishment email sequence using the product and customer reorder data below.
Use:
— normal product replenishment cycle
— each customer’s previous reorder intervals (specific to replenishment products)
— quantity or pack size
— recent browsing behavior
— whether the customer has already reordered
Recommend:
— when the first email should arrive
— whether follow-ups are needed
— what each email should say
— when the sequence should stop
Do not use one fixed delay for every customer unless the data supports it.
Data: [PASTE DATA]
Prompt 22: Build a pre-inactivity or win-back sequence
A customer who purchases every three months shouldn’t be treated as inactive after 30 days. On the other hand, someone who usually orders every two weeks may already be showing signs of declining activity after a month.
Analyze this customer purchase and engagement data and build a re-engagement strategy.
Separate customers into:
— behaving normally
— beginning to drift
— clearly inactive
Recommend a different intervention for each group.
Use the customer’s normal purchase cycle rather than one universal inactivity threshold.
Do not automatically start with a discount.
For each group, recommend:
— timing
— message
— product/category focus
— whether an incentive is justified
— exit condition
Data: [PASTE DATA]
Identify customers who are beginning to drift or have become inactive and recommend the right re-engagement approach.
Analyze performance
Writing and sending a campaign is only part of the job. Understanding what happened afterward helps determine whether the same approach should be repeated.
Prompt 23: Analyze campaign performance and decide what should change next
A campaign with a high open rate hasn’t necessarily achieved its business objective. You need to understand whether it generated purchases, which customer groups responded, and whether the products or offers performed as expected.
Analyze the performance of this ecommerce email campaign.
I will provide:
— campaign goal
— audience
— email performance
— purchase data
— revenue
— product performance
— any available segment-level results
Do not evaluate the campaign only on opens and clicks.
Tell me:
— whether the campaign achieved its actual business objective
— which audience groups performed differently
— whether the product, offer, timing, or message appears to have influenced the result
— what should be repeated
— what should change
— what should be tested next
If the data does not support a conclusion, say so rather than guessing.
Campaign data: [PASTE DATA]
Analyze campaign results to understand what worked, what needs improvement, and what to test next.
Getting better results from ChatGPT for ecommerce email marketing
The most useful ChatGPT prompts for ecommerce email marketing are the ones that help you make a specific decision using relevant data.
If you’re looking for new campaign opportunities, provide purchase history, product information, and customer activity. If you’re choosing an offer, include discount history, margins, and previous campaign results. If you’re writing an email, provide the audience, campaign objective, product details, and examples of your brand’s voice.
You don’t need to use all 23 prompts for every campaign. Start with the decision you’re trying to make, provide the data needed to make it, and review the recommendation before putting it into action.
ChatGPT can help you work through the analysis and preparation, but the quality of the final campaign still depends on the data you provide and the decisions you make from it.
For the workflow these prompts plug into — what to connect, what to check and where a general-purpose model stops being enough — read how to use AI for email marketing and building an AI tool stack.
Nexie is an AI-native ESP for e-commerce brands. It reads your customer data across engagement, recency, category interest and lifetime value together, surfaces the audiences worth campaigning to, and builds those campaigns for your team to review and approve, inside the guardrails and marketing policies you set. See how it works →

