An AI survey that
interviews everyone.

Ask a question. ScatterSpoke follows up with every person, and
hands you the real answer.

app.scatterspoke.com/surveys/thursday-summit
JJordan's workspace
Feedback
Surveys
Retros
Question sets
AI
MCP
Agent Results Brief Final 1,214 of 10,412
Summary

Content landed; logistics didn't. Attendees rated sessions warmly but lost time to lines and overruns, and many say that decides whether they come back.

On session length, two groups disagree on the root cause.
Speakers ran over · 212
Schedule too tight · 148
What to fix before next year
1Stagger breaks by hall and double the coffee stations.
2Hard-stop keynotes; move overflow to a recorded Q&A.
3Add hosted networking tables to every break.
What people said · Break logistics
“Missed the start of three talks queuing.”
“Skipped the 2pm to beat the line.”
“One station for 10,000 people.”
Themes
Break logistics612
Keynote overruns360
Networking wanted structure241
Badge pickup97
Sentiment
61% positive24% mixed15% negative
Real answers for people at
VanguardMLBDisneyUberAirbnbHome DepotGitHubCVS HealthHubSpotAdobe
How it works

Say what you need to know.
Get the answer back.

Most surveys collect numbers or vague sentences. ScatterSpoke runs a real conversation with every person, then does the synthesis you'd otherwise spend a week on.

01 · Ask

Say what you're trying to figure out.

Type it the way you'd say it to a colleague. Spoke, the ScatterSpoke interviewer, turns it into a goal, an audience and the topics worth covering.

What do you want to learn, and who's answering?

Why customers who cancelled this quarter left, and what would have kept them.

Drafted the spec. Should I ask what they switched to, or keep it to their experience with us?

Live survey spec
Goal
Why customers cancelled, from the customers.
Audience
Every account that cancelled this quarter
Topics
What they were trying to do
The moment they decided
02 · Conversations

Every person gets interviewed, not surveyed.

Spoke asks, listens, and follows up on what each person actually said. Each person gets a link by email or QR code and chats for the time you set. No account needed. Anonymous on paid plans.

Spoke
Named · 5 min
Take me back to the last thing you were trying to do before you cancelled.
Connecting our data source. I got halfway through the setup and it wanted a key I didn't have.
A key you didn't have. What happened after that?
Nothing, really. The trial had a week left and I wasn't going to chase our IT lead for it.
03 · Answer

Read the answer,
not the transcripts.

Sentiment on every reply
Scored answer by answer, not averaged into one number.
Themes ranked by weight
How many people raised each one, with the quotes behind it.
Contradictions surfaced
When two groups explain the same problem differently, you see both.
Agent Results Brief Final
Summary

Price barely comes up once Spoke follows up. Most churned customers describe the same wall: a data-source setup step that needs a credential they don't own, with the trial clock running. Several say they never saw the product working with their own data.

On what would have kept them, two groups disagree.
A sample dataset to try · 17
Someone to set it up · 9
What to fix before the next cancellation
1Ship a sample dataset so the product works before any credential.
2Pause the trial clock until the first data source connects.
3Send the credential request to the account's admin, not the trial user.
What people said · Setup needs a credential
“Setup asked for admin access I don't have.”
“I never saw my own numbers in it.”
“The trial ran out before the setup did.”
Themes · 31 responses
Setup needs a credential24
Never saw own data14
Trial too short to finish10
Price4
Sentiment
16% positive45% mixed39% negative
Use cases

Whatever you're
trying to figure out.

A team, a community, ten thousand strangers, or five people who just quit. Same three steps, same kind of answer.

Engineering
Includes retro boards, free

“What's slowing the team down this sprint?”

What slowed you down most this sprint?What did you work around instead of fixing?
Review wait time, not review quality, is the bottleneck.
See the use case
Events

“10,000 attendees walked out an hour ago. What did they think?”

What's one moment from the event that stuck with you?What did you hope to get out of coming?
Content landed, logistics didn't. Fix the breaks before the sessions.
See the use case
Community

“Why did members go quiet after month two?”

What did you hope to find here when you joined?What happened the first time you asked something here?
Members leave when their first ask goes unanswered. It's onboarding, not content.
See the use case
Product

“Why did these 40 customers churn?”

What's the main reason you cancelled?What were you trying to do when you left?
They didn't churn on price. They churned on a setup step nobody finishes.
See the use case
Marketing

“Why did they click but not sign up?”

What did you expect to see when you clicked?How did the page compare with what the ad said?
The ad promised a free guide. The page asked for a credit card.
See the use case
People

“How are our new hires really doing at 30, 60 and 90 days?”

What did you need this week that you didn't have yet?How did your first week compare with what you expected?
The cohort is engaged but nobody knows who owns their onboarding.
See the use case
Start from a question

Not sure what to ask?

Every question in these guides comes with the follow-up to ask next. Copy them, or send one and let Spoke ask the follow-ups.

Agent native

Or never open
ScatterSpoke at all.

Create, send and read surveys from Claude or ChatGPT. Ask for a survey in the chat you already have open, and the answer comes back to the same thread.

Claude ChatGPT
33 tools, from creating a survey to generating the answer.
One command to connect. Works on every plan, including free.
Respondents can answer from their own assistant, too.
Claude·ScatterSpoke connected
Survey everyone who came to Thursday's summit and tell me what to fix for next year.
Done. I set up an agent survey for the 10,412 people who scanned in, with two topics: session quality, and breaks and venue logistics. Invites are going out now.
create_surveyemail_survey
Any signal yet?
1,214 responses so far, and the answer just firmed up. Short version: content landed, logistics didn't. Break lines and keynote overruns are the top two themes, and there's a split on why sessions ran long.
get_survey_resultsgenerate_results_brief
Recommended actions
1  Stagger breaks by hall and double the coffee stations.
2  Hard-stop keynotes; move overflow to a recorded Q&A.
3  Add hosted networking tables to every break.
Reply to Claude…
Free forever

Classic surveys and retro boards. Free. No limits.

Type your questions, people answer, you see the raw results. Unlimited surveys, unlimited responses, unlimited retros. And your first 10 agent surveys are on us.

Ask your first question.

Free to start. No card. Your first ten agent surveys are on us.