Your Rep AI assistant collects satisfaction feedback directly inside the chat. Shoppers can rate a conversation whenever they want to, and your AI also asks for a rating on its own when a conversation reaches its natural end. This gives you direct feedback on how shoppers feel about the help they received.

A rating starts in one of two ways:

• The shopper asks to rate. Tapping the rating icon in the top row of the chat widget opens the rating question. This is available at any point in a conversation.

• Your AI asks automatically. When a conversation reaches its end, your AI invites the shopper to rate it before signing off. This is covered in detail in the next section.

Either way, the shopper sees the same question — "Was this conversation helpful?" — with Yes and No reply options. The shopper can tap a reply option or simply type their answer, and both are recorded the same way.

Rating is always optional and never blocks the conversation. When a shopper does answer, it's logged instantly for your review.

When your AI asks "Is there anything else I can help you with?" and the shopper confirms they're all set, your AI invites them to rate the conversation as it signs off. The closing message and the rating question arrive together, so you collect feedback from shoppers who would otherwise leave without saying anything.

What the shopper sees

• Tapping Yes gets a short acknowledgement, and your AI checks whether the shopper needs anything else.

• Tapping No gets an apology and an offer to connect the shopper with your team, so an unhappy conversation has somewhere to go.

Asked once per conversation. The rating question appears at most once in a conversation. If a shopper says they're finished, keeps chatting, and reaches the closing question again, your AI won't ask a second time.

Where it applies. The automatic rating question appears in the chat widget on your store. Conversations that reach you through other channels — email tickets, your helpdesk, or Instagram, Facebook, and WhatsApp — end without it.

The wording is yours. The rating question and its Yes and No reply labels are System Messages, so you can rewrite them in your own voice from your AI's settings, and they follow your store's languages like every other system message.

Not every conversation ends with the question. It appears when a shopper confirms they don't need anything more. Conversations that end another way — the shopper closes the widget, stops replying, or gets handed to your team — finish without a rating. A shopper who ignores the question still ends the conversation cleanly, and "no rating" is an ordinary outcome rather than an error.

Homepage Dashboard: - Check "Conversations Reported as Unhelpful" in your key metrics

Support Analytics:

1. Go to Analytics > Support

2. Open the Customer satisfaction survey chart to compare helpful and not-helpful ratings across the period you've selected

Shopper Intelligence:

1. Go to Insights > Shopper Intelligence

2. View "Reported as helpful/unhelpful" metrics in the "Forms of Assistance" section

3. Use the "Unhelpful" tab to focus on negative feedback

Conversations Page:

1. Navigate to Conversations

2. Filter by Reported as helpful or Reported as unhelpful

3. Review specific conversation transcripts — the rating question and the shopper's answer appear in the transcript in sequence with the rest of the chat

Ratings collected automatically at the end of a conversation sit alongside ratings shoppers give at other moments, so these screens show a combined total rather than separating the two. The Customer satisfaction survey chart counts shoppers who gave feedback or followed an AI product recommendation, so its total is broader than the count of explicit ratings alone.

"Helpful" Ratings: Show what your AI does well - topics it handles effectively and responses that satisfy customers.

"Unhelpful" Ratings: Highlight improvement opportunities like:

- Questions your AI couldn't answer completely

- Responses that missed the customer's intent

- Unclear or confusing information

Weekly Review:

1. Check unhelpful conversation trends on your homepage

2. Review unhelpful conversations in Shopper Intelligence

3. Identify patterns in problematic responses

4. Update AI training or knowledge base accordingly

Focus on High-Impact Issues:

- Address topics with the most unhelpful ratings first

- Look for recurring problems across multiple conversations

- Track improvements over time

• Let it flow naturally: Customers will provide feedback when they want to

• Focus on patterns: Look for recurring issues rather than individual ratings

• Take action: Use feedback to make real improvements to your AI's responses

Most customers don't rate every conversation, and that's normal. The feedback you receive provides valuable insights into your AI's performance and helps you continuously improve customer experiences.

• AI Recommendations: Get specific suggestions based on unhelpful feedback

• Shopper Intelligence: Analyze feedback alongside conversation topics

• Conversations: Review transcripts of rated interactions