Live in-meeting nudges vs coaching after the meeting

The pitch is the most appealing idea in the category, and Poised is built on it. Software watches your meeting and tells you, right then, that you are talking too fast or interrupting or hedging. Fix it in the moment, instead of learning about it later.

It works less well than it sounds, and the reason is not the software.

The attention problem

In a meeting, you are already doing several demanding things at once. Composing the next sentence. Tracking whether the point landed. Following what the other person said. Watching the room. Holding the thread of your own argument.

Adding a prompt to read does not add capacity. It takes some.

The sequence is unforgiving. The nudge appears. You notice it. You read it. You work out what it wants. You try to adjust. Every one of those steps takes attention from talking, and the sentence you were in has already finished. You have received advice about a moment that no longer exists, at the cost of the moment you were in.

There is a reliable tell for this, reported by people who use these tools: the part they actually find useful is the summary afterwards.

What live prompts are good for

The honest version is not that they are worthless. It is that they work for a narrow class of signal.

A single binary indicator you can absorb without interpreting, glanced at rather than read. A pace bar. A talk-time share. Something with no thinking between seeing it and acting on it.

They degrade fast as the advice gets richer. “You are speaking quickly” is actionable at a glance. “You hedged before that number” requires you to recall what you said, agree, and rephrase, which is a conversation with yourself that you cannot have while talking to someone else.

What after works for

Afterwards, the constraint disappears. You have attention. You can look at the actual words. You can disagree.

More importantly, afterwards can see things the live view structurally cannot.

The shape of a whole exchange. Whether you answered the question that was asked or the one you wished had been asked. That you dropped a position after being pushed but before being given any new information. That a point was made twice. That the answer arrived ninety seconds in.

None of those are visible at a single moment, because they are patterns across several. A tool watching one sentence cannot see the sentence it relates to.

The trade you are making

	Live, as in Poised	After, as in Vocanti
Timing	While it matters	When you can think
Attention cost	High, at the worst moment	None
Sees	One signal, now	The whole exchange
Acts on	Simple metrics	Patterns and specific lines
Changes habits	Rarely, no repetition	Across sessions

Rehearsal tools like Yoodli and Orai sit outside this table entirely: they see neither the live moment nor the real meeting, because the speech was produced for them.

The strongest argument for after is not comfort. It is repetition. A habit changes when the same thing is named several times in situations you recognise, and a prompt you half-read in a meeting you barely remember does not accumulate into anything.

The reasonable objection

If the feedback comes afterwards, the meeting is already over. You cannot un-hedge the date you gave.

True, and it matters less than it sounds, because you are not trying to fix one meeting. You are trying to change a habit that shows up in every meeting. What matters is whether it happens less in the next thirty, and for that, the meeting you just lost is data rather than a loss.

Which is also the honest limit of both approaches. Neither one saves the conversation you already had.