Intro:
If you’ve used generative orchestration in Copilot Studio, you’ve seen the comforting “thinking” animation while the copilot works. Classic orchestration doesn’t do that out of the box. So when you kick off a longer task—like a RAG lookup or a flow call—the chat can look empty. No typing dots. Just silence.
That silence is a UX bug, not a feature. The fix is quick: send a Typing event immediately before the slow step, then add a short acknowledgement message to set expectations.
Configuration:
Locate the slow step or Topic where you have user feedback on the poor UX expereince (May be based on the user activity we might have logic where the heavy work starts: Create generative answers, a flow, or an HTTP/API call.)
Set Event type to Typing.
The example below the timer is set as 2100 ( milliseconds)
Add a brief acknowledgement message Right after the Typing event, add a normal message to set expectations. Keep it human and specific:
“Got it — searching our knowledge base. This may take a few seconds.”
“Working on your request — I’ll be right back with the details.”
Good UX patterns to include:
- Time cue: “This usually takes ~5–10 seconds.” It sets expectations and reduces anxiety.
- Specific context: “Checking SAP invoices…” feels more reassuring than generic “Working…”
- Channel-aware: Typing events are short-lived (around 2-4 seconds). Always send an acknowledgement text so users see something even if the channel ignores Typing.
Closing Notes:
A great bot isn’t just smart — it’s considerate. In classic orchestration, adding a Typing event plus a short acknowledgement transforms a “silent” wait into a clear, reassuring experience. It takes less than a minute to wire up and pays off immediately in user confidence and perceived speed.
Top comments (1)
Is
act_typea value you provided... or is that part of the framework schema? In other words, does that need to be set to that, or is the value arbitrary?Never got this to work on Teams channel, but admittedly didn't play around with it too much.