Two surveys, two reports
How you'll know: Spring and autumn NPS ran as separate surveys, and there's no single trend line.
Fix: Intercom can't combine stats from different surveys into one report. Export each survey's results and join them yourself.
Your head of support wants a satisfaction number, and product wants to know why people drop a feature. Intercom gives you three ways to ask. Conversation ratings send a five-emoji CSAT question after a teammate closes a conversation. Surveys, built under Outbound, run from a single NPS question to fifteen questions and can save each answer to the person. Fin, Intercom's AI agent, collects its own CSAT. Each has its own setup and its own report.
Choose small or large format and save. Name it after what it asks, such as "Onboarding: role and experience", so the report is easy to find later.
Pick from NPS, numeric scale, star rating, emoji rating, dropdown, short text, long text and multiple choice. One survey holds up to 15 questions, in up to three sections of five.
Watch for: branching only works from closed-ended questions, so put the free-text question after the rating that decides who sees it.
Turn on Store answer to a user attribute for every question you'll want later. Answers save each time the person submits a step, so partial responses still count.
Watch for: the attribute's data type. NPS needs a number attribute, and multiple choice answers need a text one.
Add rules for who sees it, pick web, iOS or Android under Deliver first on, then send it automatically or share its link.
Watch for: a survey on a fixed schedule pauses itself after 30 days.
Survey answers sit on each person, as the attributes you chose, and each survey has its own report page with a CSV export. Individual responses can also be pulled through Intercom's Data Export API, though aggregated figures such as response rates and NPS scores only come from the report itself.
Conversation ratings live on the conversation and count toward the teammate who handled it. The CSAT report shows each teammate's score, and its Negative remarks view lists the badly rated conversations with the customer's comments. A workflow can also wait for the rating and branch on it, so an unhappy customer gets a different reply in the moment.
How you'll know: Spring and autumn NPS ran as separate surveys, and there's no single trend line.
Fix: Intercom can't combine stats from different surveys into one report. Export each survey's results and join them yourself.
How you'll know: Most customers picked the middle face, and the score still came out under half.
Fix: Intercom's CSAT is the share of positive ratings, the two happiest faces, out of all ratings. Report the full spread next to the score so a run of neutral ratings reads as neutral.
How you'll know: Customers who write in weekly get a rating request after every closed conversation.
Fix: Tag customers you've asked recently and exclude that tag in your rating workflow.
How you'll know: Five users at one customer answered NPS, and the account owner sees none of it before the renewal call.
Fix: Roll the answers up to the company, alongside its CSAT, before they reach the account record.
Yes. Surveys, under Outbound, support NPS and seven other question types, in small or large format, on web and mobile. Intercom's pricing page lists Surveys with the Proactive Support Plus add-on rather than the base plans. Conversation ratings (CSAT) are a separate feature, switched on through Simple automations.
Inside Intercom, yes. A workflow can wait for a conversation rating and branch on it, sending a different message or routing the conversation when the score is low, on Advanced or Expert plans. Survey answers saved to attributes can drive audience rules for later messages. Acting in another system, such as a CRM account record, needs a connection built on Intercom's API.
On the person, as user attributes, when Store answer to a user attribute is on for that question. They save step by step, so partial answers stay. Each survey's report also offers a CSV export, and individual responses come out through the Data Export API.
Yes, with a build in between. If your CRM's app from the Intercom App Store already maps the attributes you need, use that first. Account-level scores are where it stops: n-frames builds the roll-up that pulls responses through the Data Export API, joins them with conversation ratings, and writes one feedback score per company to the CRM or health-score dashboard.
Yes. Fin CSAT asks for a rating on conversations Fin handled, on plans that include Fin AI Agent, and it's reported separately from teammate ratings.