What are AI-powered segments?

AI-powered segments use the Customer insights filter to find groups of shoppers from your store’s order history. You pick a shopper journey, such as Loyal shoppers, and Instant AI shows who is in it, what is happening with them now, and the revenue you could make if you email them.

Why you need one

  • Other filters describe shoppers. They do not tell you which group to email now or why.
  • Thresholds such as “at risk” or “overdue to reorder” are different for each store. Customer insights calculates them from your own orders, so you do not have to guess a number of days.
  • Some groups, such as shoppers who are due to reorder, come from a prediction. You cannot build them with the other filters.

How to set it up

Before you start

  • Your store needs order history. If there is not enough, some opportunities are disabled. See How it works below.

Building an AI-powered segment

  1. Go to Shoppers in your Instant AI dashboard
  2. Click the filter icon next to Search for an email
  3. Under AI-powered, select Customer insights
  4. Under Shopper Journey, select a journey
  5. Under Opportunities, select one or more opportunities
  6. Read the AI-powered insights panel to check the audience
  7. Click Apply
  8. Click Save
  9. Check the segment name. Instant AI suggests one from your selection, and you can change it
  10. Confirm to save

Tip: You can also use Customer insights when you choose a campaign audience, in Smart Rules and in Coupon Rules. A segment that you save on the Shoppers page shows in those places too.

How it works

Shopper journeys and opportunities

Each shopper journey has a set of opportunities. When you select more than one opportunity, a shopper who matches any of them is in the audience. The number next to each opportunity is how many shoppers match it now.

First-time shoppers

  • Recently acquired & active: placed their first order in the last 30 days, and have not started to slow down
  • New Customers: bought recently, with few orders and low spend so far, in your store’s RFM (recency, frequency and spend) ranking
  • One-time buyers: placed exactly one order
  • Acquired but not repurchased: placed one order, and it is longer than your store’s usual time to a second order

Repeat shoppers

  • Repeat buyers: placed two or more orders
  • Due to reorder in 7 days: predicted to order again in the next 7 days
  • Due to reorder in 30 days: predicted to order again in the next 30 days
  • Reorders faster than usual: orders more often than your store’s usual reorder time

Inactive shoppers

  • Overdue to reorder: past their predicted reorder date, but not yet past the point where most of your shoppers stop buying
  • Slowing down: no order for longer than the time when your shoppers usually start to slow down
  • Likely to stop buying: no order for longer than the time when most of your shoppers stop buying
  • High spenders falling behind: in your top 20% by lifetime spend and overdue to reorder

Engaged shoppers

  • Frequent buyers, large baskets: three or more orders, with an average order value more than 25% above your store’s typical order value
  • Repeat buyers, medium baskets: exactly two orders, with an average order value within 25% of your store’s typical order value
  • One-time buyers, small baskets: one order, with an order value more than 25% below your store’s typical order value

Loyal shoppers

  • Top 10% by lifetime spend: your top 10% of shoppers by total spend
  • Top 20% by lifetime spend: your top 20% of shoppers by total spend
  • Buying on schedule: two or more orders, and their last order is more recent than the point where your shoppers usually start to slow down

Opportunities with an AI chip come from a prediction model. The segment shows an AI chip in your saved segments list when it uses Customer insights.

The AI-powered insights panel

When you select a journey and opportunities, the panel shows three rows:

  • Who they are: how many shoppers are in the audience
  • Right now: a benchmark from your store, for example how many days your shoppers usually take to reorder
  • If you send: a suggested email and a revenue forecast

The forecast is the number of shoppers in the audience, multiplied by your store’s repeat purchase rate, multiplied by your typical order value. The footer shows the audience as a share of all your customers.

Note: Instant AI does not forecast revenue for an audience under 100 shoppers, because the number would move too much to be useful.

When your store does not have enough order history

Each opportunity needs a benchmark from your orders, such as your usual time to a second order or your typical order value. If your store does not have that benchmark yet, the opportunity is disabled. If no opportunity in a journey has the history it needs, the panel shows Still gathering your shopper insights.

FAQ

Do AI-powered segments update by themselves? Yes. Like all segments, they are dynamic. A campaign checks the audience again when it sends.

Can I choose the number of days for “Slowing down” or “Likely to stop buying”? No. Instant AI calculates these from your store’s order history. The Right now row shows the numbers it uses for your store.

Can I combine Customer insights with other filters? Yes. Add other filters on the Shoppers page in the usual way. For example, add Location to send a win-back email only to shoppers in one country.

Why is an opportunity greyed out? Your store does not have enough order history for that opportunity yet. Remove it from your selection to apply the filter.

For more on segments, see How to Create and Save Segments and How to Use Segments in Campaigns.