ScatterSpoke v3 (Aug. 2026)

ScatterSpoke v3 is a new platform, not an upgrade to the old one. It was rebuilt from scratch and first released in August 2026, and it changes what the product is: v2.5 collected feedback into boards you read, while v3 collects feedback into evidence you can question.

This page is the inventory — everything v3 does, grouped by what you’d use it for. If you want the argument for why it’s shaped this way, read What is ScatterSpoke instead.

Coming from v2.5

Your workspace, your people, and your history came across. Retrospective boards from the old platform were migrated with their original columns, cards, votes, and comments intact, and they open as read-only history you can reference or reopen — see Migrating legacy boards.

Three things are genuinely different, and they’re worth knowing before you go looking for the old ones:

Async feedback became Surveys. The old scheduled-feedback flows are now Surveys, which do the same job with a real builder, a participant surface, and response transcripts.
Retro phases were renamed and reduced. Reflect is now Feedback, and the manual Group phase is gone — grouping is done by AI theme clustering.
Plans were restructured. v3 has four tiers. Lapsed paid plans from v2.5 land on Free rather than being cut off. Plans at a glance has the current shape.

Asking people things

Two ways to ask, and the difference is who decides what gets asked next.

Static Surveys are authored question sets — you write them, everyone answers the same thing. Unlimited on every plan including Free. The builder covers nine question types plus an NPS preset, per-question hints, allow-skip, and accessible reordering.

Agent Surveys invert that. You describe the outcome you want in a Design interview, and the agent interviews each participant turn by turn, following what they actually said. Two people can end up in different conversations. Paid plans.

Around both:

A templates gallery with a versioned library, featured templates, and provenance recording which template a survey came from.
Distribution by link, QR code, or embed, with expirations you can change after publishing.
Recurring series that compare each run against prior themes.
A participant surface with save-and-resume, terminal states, and a branded entry gate — no account required.
Responses and transcripts, including an asynchronously generated Results Brief, a drop-off funnel, and real median completion time.

Running retrospectives

Retrospectives are live collaborative boards.

Three phases — Feedback, Vote, Review — driven from a facilitator dock, with a board lock and a shared timer.
Live presence, typing indicators, and per-phase participation nudges.
Private feedback and anonymity, including a hidden-until-reveal mode.
Guest join with identity tiers and a rotatable link — guests never consume a seat.
AI theme clustering with realtime themes, cluster chips, and a prompt when the vote phase opens. Probe deeper on a theme drafts a pre-briefed Agent Survey from it.
A temperature check — anonymous 1–5 mood with a live average.
Six board formats in a unified templates library.
A closed-retro summary with hero statistics, top-voted cards, energy bars, CSV export, Run it again, and an across-retros trends panel.
Mobile layouts below 780px.

Retro AI features are entitlement-gated.

Questioning what you’ve collected

Ask is a natural-language workspace over your accumulated feedback. You ask; it answers with evidence.

Threads with pinning, @-mentions, and a command-palette handoff.
Grounding to specific surveys or to a closed retrospective — a Talk to this retro action opens a thread already scoped to that board.
Artifacts: tables, KPI rows, donut and pie charts, segment grids, and theme breakdowns with per-theme sentiment. Charts and tables are built from captured tool results rather than re-generated prose, so what you see traces back to real rows.
Cited feedback names the survey or retro it came from.
An Ask credit pool with visible accounting.

Signals are the themes the platform detects on its own — extracted per response, aggregated across a survey, and surfaced org-wide. The catalog covers the seeded set, custom signals let you define your own, and events and evidence is the audit trail behind any given one. Conflict signals surface where a team disagrees with itself; cross-team friction surfaces in org-wide queries.

Running the workspace

Teams organise people and scope what they see: retros filter by team, members see their own teams in account settings, and admins manage the roster.

Roles and permissions with system and custom roles.
Invites and seats — Participants never consume a seat. Participants can now request one from their links hub.
Audit log, data retention, and organization settings.
Scope-restricted API keys and the MCP server, which is available on every plan with its insight tools entitled separately.

Billing runs on four tiers — Free, Pro, Business, Enterprise — with annual billing, an always-on invoice preview, coupon codes, usage packs with instant invoicing, and seat management.

Branding covers accent colour, app name, light and dark logos, branded email from a verified domain, custom domains with live SSL status, white-label options, and social preview cards on shared survey links.

Signing in

Single sign-on supports Okta over OIDC and ADFS, with domain verification and just-in-time provisioning. A workspace can be set to require SSO, in which case password sign-in and password resets are refused for governed identities.

Individual accounts support passkeys and two-factor authentication, both settable during onboarding. Password reset links last one hour. Participant SSO verification covers what applies to people answering surveys rather than entering the workspace.

Connecting other systems

The MCP server connects an AI coding agent — Claude Code, Claude Desktop, or Codex — directly to your workspace, so you can create a survey and read its results without opening the web app. It’s available on every plan, with the insight tools entitled separately.