16 August 2026
Breakout rooms fail for the same reason most of the time: the facilitator sends people off with a vague topic and no clock, then loses ten minutes trying to reassemble what six groups just decided. The button that splits a call into smaller rooms is trivial. Getting useful output back out of them is the actual problem, and it's a facilitation problem, not a software one.
Splitting a meeting into rooms is one click on every platform that has the feature. What goes wrong happens on the way back: the facilitator calls everyone into the main room, asks "so, what did your group talk about," and gets four minutes of someone trying to reconstruct a conversation from memory while three other people wait. If the point of the breakout was to surface disagreement or generate options, that reconstruction step is exactly where the specific, useful parts get flattened into "yeah, we talked about the same stuff."
The fix isn't a better recap format. It's deciding, before you split people up, what artifact each room needs to produce and how it gets back into the main record without a human being the transfer mechanism.
Under five minutes, a breakout room produces almost nothing beyond restating the prompt to each other. Over fifteen, at least one room finishes early, gets restless, and starts a side conversation that has nothing to do with the task \u2014 by the time you call them back, the thing you wanted has been talked past.
Eight to twelve minutes is the range that works for most task types: enough for one person to disagree with the framing, someone else to counter, and a rough answer to emerge. If the task genuinely needs longer \u2014 reviewing a document together, working through a design \u2014 give it twenty minutes but check in once at the midpoint by dropping into each room briefly rather than waiting for the full block to run out.
"Discuss X" is not an instruction, it's a topic. Rooms that get a topic instead of a task spend the first three minutes deciding what they're even supposed to produce, which eats the time you budgeted for actual discussion.
A workable instruction has three parts: a specific question, a named output, and who's responsible for bringing it back. "Decide whether we ship the redesign this sprint or next \u2014 one sentence, plus your top reason \u2014 and whoever's screen is on top when time's up reports it" gives a room nothing to negotiate about except the actual question. Write it down somewhere everyone in every room can see it, not just in the main-room chat they left behind.
The report-back format matters more than most facilitators plan for. A round of "someone from each room summarizes" is honest but slow \u2014 with five rooms at two minutes each, that's ten minutes of the main meeting spent on recall instead of decisions, and the summaries compress unevenly depending on who happens to talk.
A better default: have each room drop their answer into a shared doc, chat thread, or notes tool while they're still in the room, not after. Then the facilitator reads the outputs aloud or screen-shares them rather than asking people to re-narrate from memory. This turns the return from a series of live retellings into a five-second read of what's already written down \u2014 the room's actual words survive instead of someone's summary of them.
Most video platforms treat a breakout room as a genuinely separate call: separate audio stream, and on some tools, a separate recording or transcript file that never gets linked back to the meeting it came from. If your record-keeping is per-meeting, a breakout discussion that produced the actual decision ends up orphaned in a file nobody thinks to open three weeks later when someone asks "why did we go with option B?"
AVAY doesn't currently split a call into native breakout rooms \u2014 if you need that, you're running it on the platform that supports it and pulling the outcome back into AVAY's notes yourself, which is a real limitation worth planning around. What it does do is keep a running transcript and decision log attached to whichever meeting is open, and its search reaches across every past meeting, so if a breakout ran as its own AVAY call, someone can later search for what was said in it without hunting for a separate file. That's not the same as a room being automatically stitched into the parent transcript \u2014 it's a way to find the room afterward, not a substitute for merging it at the time.
| Time cost | What gets lost | What survives afterward | |
|---|---|---|---|
| Verbal report-back to the group | 5\u201310 min for 5 rooms | Nuance, disagreement, exact wording | Nothing unless someone was typing notes live |
| Shared doc or chat, written during the room | Near zero, happens in parallel | Depends on who writes and how fast | A written line per room, in the room's own words |
| Each room recorded as its own call, transcript linked back | None extra if transcription runs automatically | Little, if the room stayed on task | A searchable, attributable record per room |
Keep each room at three to five people \u2014 below three, there's no real discussion, and above five, at least one person goes quiet. For a group of twenty, that's four to six rooms, which also sets your report-back time: budget roughly one minute per room if you're reading written outputs, more if people are talking.
Only if a specific room needs unblocking, and briefly. A facilitator who parks in one room for the whole segment can't check the others, and their presence changes what a group is willing to say. Drop in, ask one question if something's off track, and leave.
Give them a follow-up question tied to the same task rather than letting them sit idle or wander into an unrelated conversation \u2014 \"if you have time left, note the one thing you disagreed on\" keeps the extra minutes useful instead of wasted.
If the discussion could affect a decision you'll need to justify later, yes \u2014 a written summary from memory loses the reasoning, and the reasoning is usually what someone asks for afterward. If the room is just generating quick reactions, a one-line written output is enough and recording adds nothing.
Cap it structurally: require each room to produce one sentence, not a summary, and read those sentences aloud yourself rather than handing the floor to each room's spokesperson. A round where five people each talk for ninety seconds costs more time than the breakout itself.
A breakout room is only as useful as the instruction you gave it and the record you pull out of it \u2014 fix the task, the clock, and where the output lands, and the button underneath barely matters.
Meetings that take their own notes, in the browser: avay.ai.