The answers are in. Now what?
Signals reads what comes back — summarizing open text, surfacing findings, and suggesting the next question to ask, the audience to build, and the campaign to send.
Reading responses is work. Signals does the first pass.
You still make the decisions. Signals removes the hours between the responses arriving and the decision being obvious.
AI summaries
Open-text answers summarized into readable themes instead of a spreadsheet you never open.
AI-generated findings
Patterns across responses, customers, and orders, surfaced as findings worth investigating.
Suggested next questions
What the current survey did not answer, phrased as the question to ask next.
Suggested audiences
The groups your responses already support, proposed as ready-to-build audiences.
Suggested campaigns
A recommended message that follows directly from what a finding shows.
Outcome awareness
Findings account for what customers did after answering, not only what they said.
Collect, summarize, surface, suggest.
Response intelligence runs continuously as responses arrive, not as a report you request once a quarter.
-
1. Collect
Responses arrive from surveys and Sparks across every surface.
-
2. Summarize
AI summaries turn open text and answer spread into readable themes.
-
3. Surface
Findings highlight the patterns that connect answers to outcomes.
-
4. Suggest
Signals proposes the next question, audience, and campaign.
The most valuable answers are the hardest to read.
Open-text responses are where customers say the thing you did not think to ask about. They are also the answers most likely to go unread. AI summaries make them usable.
Explore survey analytics- Themes extracted from free-text answers, with the underlying responses still readable.
- Sentiment and recurring language grouped rather than counted.
- Summaries that keep the customer’s own phrasing where it matters.
- A shortcut through hundreds of responses — not a replacement for reading them.
What a generated finding looks like.
An illustrative example of the format — a pattern, the evidence behind it, and a suggested next step.
Illustrative example
Customers who identify “daily focus” as their main goal are 2.3× more likely to reorder Lion’s Mane within 45 days.
- Suggested next question — ask focus customers when they take it
- Suggested audience — focus goal, no reorder after 45 days
- Suggested campaign — replenishment reminder timed to the routine
Useful, and honest about its limits.
AI is good at reading a thousand responses quickly. It is not a substitute for your judgement about your own customers, and Signals is built that way.
- Summaries and findings are generated from your own response data.
- Every finding points back to the responses behind it.
- Findings are a prompt to investigate, not proof of causation.
- Suggestions are proposals — nothing is published or sent without you.
- Any example shown on this website is an illustrative example, clearly labelled.
Signals is a complete Shopify product on its own. Compass adds cross-product intelligence on top, and is never a prerequisite.
Start learning directly from your customers.
Build your first survey, publish a Spark, and turn every response into usable customer intelligence.