Typed Messages
Every message carries a type — decision, question, task, approval — and the type does the work.
Inside Typed Messages
Aanty’s unit of communication isn’t a plain string in a flat channel. Every message is AI-assigned, and user-overridable, one of a fixed set of types, and that type drives routing, notification priority, search facets and analytics.
A real substrate, not a label
message, question, answer, decision, update, task, approval, alert, doc, event, celebration — each behaves differently downstream.
Classification never blocks send
AI assigns type after the message is persisted — sending is always instant, never gated on a model call.
User-overridable
AI gets it right most of the time; you can always correct the type, and the correction is what the model learns from.
The detail
Typed messages are the substrate everything else in aanty builds on: routing, notification priority, search facets and analytics all key off the type. Classification is applied as post-send enrichment, so it never sits on the critical path of sending — and a human correction to the type is what keeps the classifier accurate over time.
Keep exploring
Where this connects — in the product and in the rest of the catalog.
See Typed Messages on your own workspace
Bring a real channel. In fifteen minutes we’ll show you typed messages working on your team’s actual conversations, not a demo script.