Why This Mattered. Loopwise’s support team was spending most of its day on the same handful of question types — password resets, billing changes, integration errors — leaving little time for the complex cases that actually needed a human. At the same time, new customers were dropping off during onboarding because setup steps that should have taken minutes were taking days to get a response on. The team needed automation they could trust with live customer conversations, not a chatbot bolted on as an afterthought.

What We Did. Centos Digital built a support copilot on top of Loopwise’s existing helpdesk, using a retrieval layer over the product’s documentation and past resolved tickets so responses stayed grounded in real answers rather than guesses. We designed a tiered handoff: the AI agent resolves common requests directly, drafts responses for the trickier ones for a human to approve, and escalates anything sensitive straight to the team, with full visibility into why each decision was made.

On the onboarding side, we replaced the static setup checklist with an in-product agent that walks new accounts through integration steps, detects where a customer is stuck from their actual usage data, and nudges or auto-completes steps where it safely can. Everything was rolled out gradually behind feature flags, with the support team reviewing agent decisions daily during the first month before autonomy was increased.

The Result. Within the first quarter, the copilot was resolving the majority of tier-1 tickets without human involvement, and the onboarding agent cut the average time to a customer’s first completed project significantly. Support staff shifted from repetitive triage to handling the harder, higher-value conversations — and customer satisfaction scores on support interactions held steady even as ticket volume kept climbing