Most customer feedback programmes have the same flaw: they collect scores that nobody acts on. A survey goes out, responses land in a tool outside Salesforce, a dashboard updates, and the customer who gave a low score hears nothing. The process feels active but the loop is open. Automating CSAT properly means closing that loop — making feedback trigger action, in the same system where service happens.

Start with the trigger, not the survey

The most common mistake is to think about the survey first. Think about the event instead. CSAT should be triggered by something meaningful in Service Cloud — a case closing, an interaction completing — and timed to when the experience is fresh. When the trigger is a real service event rather than a batch send, the feedback is relevant and the response rate reflects the actual interaction.

Send through Marketing Cloud, but keep the data in Salesforce

Marketing Cloud is the right tool to deliver the journey: timing, channel, personalisation, and consent handling are its strengths. The mistake is letting the response live there. The survey response needs to come back into Salesforce, attached to the originating case and customer, so it sits alongside the interaction it measures. That requires a clean Data Extension model and a ContactKey strategy that aligns with your customer profile — not an afterthought.

Make low scores do something

A closed loop means a low score creates work. Route detractors into a service-recovery case automatically, with enough context for a human to follow up well. This is the step most programmes skip, and it is the one that turns measurement into improvement. The customer who scored you poorly should experience a response, not silence.

Resolve identity so feedback attaches correctly

If identity resolution is weak, feedback attaches to the wrong profile or none at all, and your CSAT data quietly degrades. Linking messaging sessions, cases, and survey responses to a single resolved customer is what keeps the programme trustworthy over time.

The result is a system, not a survey

Done this way, CSAT stops being a disconnected survey tool and becomes part of the service operating model: events trigger journeys, journeys capture responses, responses return to Salesforce, and low scores create follow-up. The feedback loop is closed, the data is trustworthy, and the programme actually improves the experience it measures.

This is the connected-operations pattern applied to feedback — and it is far more valuable than another dashboard.