Overview

ScatterSpoke is an AI feedback platform that turns retrospectives, standups and surveys into themes, sentiment and prioritised action items — and, since ScatterSpoke 3.0, exposes the whole survey lifecycle to AI agents over an official MCP server.

Best for: Engineering orgs in Jira/Slack that want AI synthesis across many retros — or agent builders who need a survey backend over MCP.

How ScatterSpoke describes itself

The homepage still says "the AI copilot for team feedback", aimed at agile coaches and engineering leaders drowning in retro output. The pricing page tells a different story: it is now headed ScatterSpoke 3.0 and sold in seats, "Agent Surveys" and "Ask credits" — retrospectives are not named in a single tier.

"Stop building survey plumbing. Your agents need to ask real people questions — give them a tool, not a six-month backend project."

That is the new agent-native page, pitching ScatterSpoke as survey infrastructure your agents call over MCP. The retro product is intact underneath — 20+ formats, guided reflect/group/vote/discuss phases, async collection via Jira and Slack — but the pitch has moved to AI interviews and agent tooling.

Our take

The AI here is the real thing, and it got deeper this quarter. Theme extraction, issue detection, sentiment over time and an Engagement Score are all documented, and 3.0 adds AI-moderated adaptive interviews that generate their own follow-up questions, plus natural-language querying over a pre-computed context graph. ScatterSpoke also now runs an official MCP server — live, OAuth-protected, six named tools, included on every tier down to Free.

The cost story got much worse. Business is $149/mo for five seats and $35/seat after that, so 3 teams of 8 lands at $814/mo, and the free tier shrank from 10 users to a single seat — you cannot run a team retro on it any more. There is still no dedicated health-check model (weekly temperature checks are just recurring async schedules), and integrations stop at Jira, Slack and Teams. The help-centre release notes haven't moved since September 2024, even as the marketing site shipped three new product surfaces.

Best fit: engineering orgs that want AI to do the synthesis, or agent builders who need a survey backend — if per-seat pricing isn't the constraint.

Features

Retrospectives

TemplatesCustom templatesAnonymous inputRecurring retrosSchedule future retros

Health checks

Health checks· Via recurring async feedback schedules (documented as weekly temperature checks); no dedicated health-check model

Facilitation

Meeting timerAction itemsLink actions to ideasParking lot@mentionsDrag-and-drop groupingCarry over open actionsAsync mode

Voting

Independent dot votingCustom vote count

Engagement

Polling· Via the survey builder and quantitative async questions, not an in-retro pollGIFs

AI

AI grouping / clusteringAI action itemsAI summaryAI sentiment / heatmapsAI retro / health-check insightsAI discussion prompts· AI generates async feedback questions and adaptive interview follow-ups

Reporting

Cross-team reportingOrg health dashboardAction tracker dashboardExport formats· PDF, CSV, Email, Jira, Confluence

Enterprise

SSO / SAML· Paid add-on on Business; SSO/SAML included on Enterprise (Okta and Microsoft ADFS documented)SCIM provisioning· Enterprise tier onlySOC 2 Type 2· SOC 2 Type II since launch, with annual penetration testingGDPR compliantAudit logs· Enterprise tier onlyFacilitator-only roleMeeting historyUptime SLA / dedicated support· Priority support on Business, dedicated CSM on Enterprise; no uptime figure published on the 3.0 pricing pagePublic API· Business tier and upWebhooks· Via the MCP export_results tool (push results to a warehouse, Sheets or a webhook); no webhook docs in the knowledge baseMCP server

ScatterSpoke

Overview

How ScatterSpoke describes itself

Our take

Score breakdown

Pros

Cons