After the retro, find out why. Spoke interviews every engineer on their own and follows up on the sticky notes that just said “reviews” or “CI,” so you plan the next sprint with the reason in hand.
Review wait time, not review quality, is the bottleneck.
An example of the line the Agent Results Brief leads with.Under every question is the follow-up Spoke asks when the first answer is vague. That second question is where the real answer usually is.
What slowed you down most this sprint?
Think of the last time it happened. What were you working on?
What did you work around instead of fixing?
What made working around it feel faster than fixing it?
What took longer than you expected this sprint?
Where did the extra time go?
What would have made this sprint feel easier?
Which ticket made you wish for that most?
Where did you wait on someone this sprint?
How long did it sit before anyone picked it up?
Which handoff hurt most?
What was missing when it reached you?
What did you need to start a ticket that you didn't have?
Who had it?
If one meeting disappeared, which one?
What would you do with that time instead?
What pulled you off planned work this sprint?
How long did it take to get back into what you were doing?
How much uninterrupted time did you get on a typical day?
What broke it up most often?
What in the build or deploy slowed you down this sprint?
Which step were you waiting on the last time it was slow?
Which tool do you work around most?
What does the workaround cost you in a normal week?
What part of testing a change feels harder than it should?
What did you do the last time to get around it?
What went well this sprint that we should keep?
Who or what made that work?
Who helped you most this sprint?
What did they do that made the difference?
What did you ship this sprint that you're glad about?
What made that one go more smoothly than the rest?
Every person gets interviewed, not surveyed. Spoke asks, listens, and follows up on what that person actually said. Here's an example.
Read the answer, not the transcripts: themes ranked by how many people raised them, the quotes behind each, and the places where two groups disagree.
The team is not blocked on review quality; it is blocked on review wait time. Most people describe PRs idling for a day or more, then getting a quick approve, and several say they started batching changes to avoid the wait.
An example Agent Results Brief. The counts, themes and quotes are illustrative, not from a real survey.
Keep the retro. Run it on a retro board, open with one of the icebreaker questions, and collect the sticky notes. Then send one link, and Spoke asks each engineer about the notes the board left vague.
Retro notes are short on purpose. Here is what one follow-up turns them into. Examples, not real respondents.
Not slow CI in general: one flaky suite with no owner.
One engineer is the only path to an old system, and support routes around the on-call rota to reach them.
The ticket wasn't vague; the design lived somewhere the engineer couldn't see.
What do you want to learn, and who's answering?
Drafted the spec. Want me to ask about on-call separately, or let it come up?
It follows one. Run the retro on a board as usual, then send one link: Spoke interviews every engineer, on their own, about what the board left vague.
Start with what slowed people down most and what they worked around instead of fixing. Then add the group that matches the sticky notes: handoffs, meetings, or tooling and CI.
Yes, on paid plans. Anonymous responses are a setting you choose per survey; if it isn't turned on, don't promise anonymity.
You set the time: 5, 10 or 15 minutes, 10 by default. Spoke follows up on what that person actually said, so nobody answers questions that don't apply to them.
The Agent Results Brief: a summary, themes ranked by how many people raised them, the places where two groups disagree, and the quotes behind each. Every transcript is there too.
Free to start. No card. Your first ten agent surveys are on us.
