REP AI shows why Shopify shoppers leave without buying in its Shopper Intelligence dashboard. The Visitor Drop-Off Reasons (Beta) card uses AI to sort what shoppers said in real conversations into the reasons they did not complete a purchase, such as product availability issues, a complicated exchange process or a lack of product information.
To open it, go to AI Insights > Shopper Intelligence in the REP AI Console. Other reports show the questions your AI could not answer, the topics that send shoppers to your team, and the moments when the AI steps in before a shopper leaves.
How does REP AI find out why a shopper left?
REP AI reads the conversations shoppers have with your AI and categorizes the problems they express. The Visitor Drop-Off Reasons card is built from what shoppers actually said, so each reason links back to real conversations you can read.
What does the Visitor Drop-Off Reasons card show?
The card shows the customer-expressed problems that kept shoppers from converting, as a breakdown by issue. To use it:
- Open AI Insights > Shopper Intelligence and choose a date range with the date filter in the upper right corner.
- Review the most common drop-off reasons in the chart.
- Check the conversation count and the share of drop-off conversations for each issue.
- Click a reason, either on the chart or on the problem name, to open a filtered list of the conversations where shoppers raised it.
- Fix the biggest issue first, then check the card again over a comparable date range.
Visitor Drop-Off Reasons is labeled Beta in the Console. Typical fixes include adding product information, clarifying a policy or simplifying a process such as exchanges.
Which other Shopper Intelligence reports explain lost sales?
The same page has four more reports that help you read the drop-off reasons in context:
- Forms of Assistance: the topics your AI handles, with conversation counts, the share of conversations and helpful or unhelpful votes. Tabs filter the list to All, Unhelpful, Sales or Support. One conversation can carry several topics, so topic counts can add up to more than your conversation total.
- Post-purchase vs. pre-purchase: how your website conversations divide between shopping questions and questions about an existing order. Review pre-purchase conversations to find missing product information or buying guidance.
- Shopper Emotion Analysis (Beta): the starting tone of shoppers when they open a conversation, with tags such as "Seeking Clarity," "Needs Assistance" and "High Interest."
- Top 10 Products Added to Cart by the Concierge and Top 10 PDP Redirects by the Concierge: which products convert through AI conversations and which ones draw interest but may need more information or an incentive.
For a full walkthrough of the page, see Understanding Shopper Intelligence.
How do I find the questions my AI could not answer?
Use the Missing Information page. It collects real shopper questions that your AI could not answer because the answer was not in your connected knowledge sources. In the Console, open Insights and click Missing information.
- The Open tab lists unanswered questions. The Resolved tab lists questions that now have answers.
- Each entry shows the Question, where it Comes From (such as a product page, cart or collection page), its Topic, how many Times Asked and when it was Created At.
- Click Answer, write a complete answer, optionally link it to product pages, and click Publish.
- Use the Topic filter and the Minimum Frequency filter (1 to 5) to start with questions that many shoppers asked.
See Identifying and Resolving Customer Questions with Missing Information for the full steps.
How does REP AI act when it sees a shopper about to leave?
REP AI does not wait for the shopper to ask for help. Its AI Drop-off Detection predicts when a visitor is likely to leave without buying. When the prediction reaches your configured threshold, it starts a conversation with the shopper. The Rescue Algorithm looks at behavior such as which page the shopper is on, how long they stay, whether they click or scroll, and movement that suggests they are about to exit.
When Sales Skills are enabled, REP AI approaches the shopper with a message tailored to where they are in their shopping journey. You control how proactive it is with the eagerness setting. Read more in AI Drop-off Detection Engine and how REP AI identifies and rescues disengaged shoppers.
How do I see which topics send shoppers to my team?
Open AI Support > Support Analytics and find the Escalations by Topic card. Each row shows the Topic, the number of Escalations, the number of Conversations and the Escalation rate. For example, 8 escalations in 10 conversations gives an escalation rate of 80.0 percent. A Low volume label means the rate is based on fewer than 20 conversations.
Sort by Escalations to find the subjects that create the most work for your team. Open a topic, read the conversations, then add the missing answer to your AI's knowledge or review the support skill for that request. Details are in Understanding Escalations by Topic.
Can I ask Claude or ChatGPT why shoppers get stuck?
Yes. The REP AI MCP server connects your store data to Claude and to ChatGPT. Once connected, you can ask in plain language, for example "Where do shoppers get stuck on my store?" The customer problems it reports include issues such as high shipping costs or missing product information. You can then ask for example conversations behind each problem.
A store admin connects it, and you choose what the assistant may read. Topics and customer problems update once a day, so today's conversations appear the next day. See the setup guide for Claude and the REP AI MCP server page.
Frequently asked questions
Is Visitor Drop-Off Reasons a final feature?
It is labeled Beta in the Console, along with Shopper Emotion Analysis.
What can I do about a drop-off reason?
Open the conversations behind the reason, then update product pages, add answers on the Missing Information page, simplify navigation or address the concern in your policies. Check the card again later to see whether the reason shrinks.
Why do topic counts add up to more than my total conversations?
One conversation can cover several topics. A shopper who asks about product availability and your return policy is counted under both topics, so the sum can exceed the conversation total. This is expected.
How fresh is the data?
Shopper Intelligence updates daily. New conversations do not appear immediately, and some charts use the latest full day, which is yesterday.
Do these reports include email and social conversations?
Some Shopper Intelligence metrics track website conversations only, not email. Use the same date range when you compare charts, and check each report's description for its scope.