Choosing a turnstile is a technical decision that most buyers make once. The options differ by throughput, by how they handle an emergency exit, by whether they sit indoors or outdoors — and a catalogue page rarely makes that legible to someone seeing it for the first time.
The Chatterence installation that went live this week — on tornellicontrolloaccessi.it, the access-control catalogue — works on exactly that problem: it helps visitors narrow down turnstiles and automatic gates against their own constraints, turning a list of models into a guided path.
It is integrated with the client’s existing Crisp setup rather than sitting beside it. Human escalation is part of the design, not a fallback bolted on afterwards, and an in-chat contact form collects the requests that are already qualified — so a person receives a conversation, not a cold form.
As with the first installation, the project includes an isolated business intelligence dashboard: conversations, recurring questions, user intent and the commercial signals worth acting on, in a dedicated environment.
This is the role we expect an AI web assistant to play. Not replacing people and not decorating a website, but filtering the simple requests, improving first-level guidance, and letting human operators concentrate on the conversations that need real expertise.
The implementation was carried out for Infordata Sistemi, with Marko Petelin as the reference on the client side.
