If you found Clarm looking for AI inbound chat — a widget that captures website visitors, qualifies buyer intent, and routes the good conversations — here is the update. That product still exists and still works. It is now one workflow on a broader agent builder, sharing the same machinery as everything else Clarm runs. This piece explains what changed and what did not, for readers who landed on our older inbound material.
What did not change
The inbound chat is still here. It still answers visitor questions in the moment, grounded in your own product docs, pricing, and FAQs, and still flags buyer intent so the right conversations reach the right person. One healthcare customer saw 6.1x more inbound conversations on the same traffic with it. If that is what you need, you can still run exactly that.
What changed
The chat used to be the whole product. Now it is one consumer of a shared substrate. To make the chat trustworthy in regulated settings, we built a memory layer that grounds answers in your data, source citations on every reply, tenant isolation, an audit trail, and model portability. That machinery turned out to power any governed workflow, not only chat. So the inbound chat became one of several things the builder can do.
Alongside the chat that answers live questions sit agent workflows that draft scheduled or triggered work for a human to approve: client review packs, weekly allocation packs, support deflection, compliance documentation. Same memory, same citations, same governance underneath both.
What this means if you came for inbound
You lose nothing and gain optionality. Start with the inbound chat if that is the job today. When your team later asks whether the same tool can draft the follow-up, run the weekly report, or screen new customers, the answer is yes, on the same data and the same governance you already approved. The chat was the first workflow; it does not have to be the last.
Where to read next
For the full story of the shift, read Why We Moved From Inbound Sales to an AI Agent Builder. For why owning the machinery beats owning one workflow, read From One Workflow to the Machinery That Builds Workflows. To see it on your own data, book a pilot discussion.