Skip to main content
This page tracks the most recent updates to Great Question. New releases are posted here regularly.
Stay in the loop: Click the RSS icon above to get new releases in a feed reader (like Feedly), or forward the feed into your team’s Slack or email.
​
AI moderated studies, booking webhooks, and a combined spend report

​
New

  • You can now launch AI moderated interviews and AI moderated prototype tests (beta) from the study creation page, letting an AI moderator run sessions and follow-ups for you. See AI moderated interviews and AI moderated prototype testing. (Unmoderated Surveys & Tests)
  • Your integrations can now subscribe to a participation.booked webhook and read the assigned moderator from the participant API, so external systems know about a booking the moment it happens. See the Great Question API. (Integrations & API)

​
Improved

  • Eligibility rules for invite count, participation rate, response rate, and no-show count now accept a Custom period, so you can define the exact lookback window that applies. See Candidate eligibility settings. (CRM Candidates & Attributes)
  • The MCP server now exposes richer repository consumption tools, so AI clients can search, list, and read sessions, transcripts, highlights, insights, and reels through a single connection. See MCP available tools. (Integrations & API)
  • Admins can now set an incentive wallet low balance threshold and get email alerts when the available balance drops below it, so top-ups happen before studies stall. See Incentives 101. (Billing Subscriptions & Credits)
  • The participant table for AI moderated studies now shows recording status badges and refreshes automatically while recordings process, so you can tell an import in progress from a failure. (Unmoderated Surveys & Tests)

​
Fixed

  • Self-scheduling no longer creates a duplicate participation, so a stale No Show can’t linger next to a completed session. (Studies Participants & Screeners)
  • Closing a study now refunds only the recruitment credits you were invoiced for, keeping your wallet balance accurate. (Billing Subscriptions & Credits)
  • The interview room now disables Edit booking on sessions managed by external recruitment, with a tooltip explaining why. (Studies External Recruitment & Incentives)
  • Transcript deep-links now open the session side panel even if you previously collapsed it. (Repository Search Tags & Highlights)
  • Publicly shared sessions now play back recordings that previously appeared missing. (Interviews Livestreams & Observer Rooms)
  • Rescheduling a session with a livestream no longer fails on a malformed stream URL. (Interviews Livestreams & Observer Rooms)
  • Highlight reels that mix video and audio-only clips now process and play back reliably. See Reels. (Repository Search Tags & Highlights)
  • A failed recording re-import no longer removes the existing recording; the original stays playable and the import retries. (Interviews Livestreams & Observer Rooms)
  • Manually entered interview locations now apply correctly on studies that collect participant addresses. (Interviews Scheduling & Calendar)
  • Candidates now stay in your chosen order when eligibility filters are applied. (CRM Candidates & Attributes)
  • Creating a study from a team’s page now always assigns it to that team. (Account Users & Roles)
​
Auto-complete every new study by default

​
Auto-complete every new study by default

Auto-complete can now be turned on once for the whole workspace instead of study by study. Under Settings > Governance > Permissions, the Incentive Automation toggle applies auto-complete to studies created from then on.It applies going forward only: existing studies and past sessions are untouched, and you can still override the setting on any individual study from its Automation tab. Duplicated studies keep the setting from the study they were copied from, rather than taking the workspace default.
Auto-complete acts on a session’s booked start time, not on whether the session happened. Any session still marked as Scheduled 24 hours later is completed, the thank-you is sent, and the incentive is released — even if nobody attended. To stop that for a specific participant, mark them as a no-show before the 24 hours are up, which cancels their incentive. Leave the default off if you’d rather review attendance before paying.
  • Auto-complete is now described consistently across workspace settings, study automation, and permissions: each one states that the 24-hour window starts at the booked interview time and that the incentive releases regardless of attendance.
​

​
MCP Skills

Admins can now author reusable instructions, skills, that every AI tool connected through the Great Question MCP follows automatically. A skill captures how your team does something (screener conventions, incentive rules, grounding standards for insights), so agents follow your process instead of guessing at it.Each enabled skill is advertised to connected agents with its name and description, and the agent loads the full instructions for whichever ones are relevant. Skills can also be set as prerequisites for specific MCP tools: when an agent calls a gated tool without the required skill loaded, the call is denied with a message pointing at the get_skill tool to load it. Manage skills under Settings → AI & MCP.See MCP: Skills.

​
Richer MCP context for agents

Agents connected via MCP can now pull a structured participation event log summarizing what each participant said and did in a session, giving them richer context for analysis and follow-up. See MCP available tools.
  • The incentives page now includes a sortable, filterable table with column header controls, making it faster to find and manage payouts across a large history. See incentives 101.
​
Autonomous MCP agents
Agents can now connect to Great Question on their own, give an agent a secure API token once, and it connects in the background whenever it needs to, without a human driving it. It only sees what that user is allowed to see, the same permissions and privacy as a normal login.Agents can also start from your saved study templates instead of rebuilding every screener question by hand, with new tools to list, search, read, and apply templates.
  • Editing an account user’s role now surfaces the exact reason a change is blocked (seat limits, permissions), so you can resolve it without guessing.
  • When Microsoft rejects a calendar event because the connected mailbox is over quota, the calendar owner and study owner now see a clear in-app warning naming the calendar, with a link to the interview.
  • Adding interview moderators by email now works when the person is already an account member.
  • Creating a study via MCP without a team now assigns your default team automatically.
​
Great Question MCP is here
Great Question now runs an MCP server, giving an AI agent, in Claude Desktop, Claude Code, ChatGPT, Cursor, or any other MCP-compatible tool, read and write access to your studies, sessions, transcripts, highlights, insights, and participant data. It can manage candidates, build studies, and handle recruitment and scheduling directly from a chat conversation.Connect at https://greatquestion.co/api/mcp/v1. Every path uses OAuth 2.1 with PKCE, so you authenticate through your browser with your existing Great Question credentials, no API keys or secrets to manage. See MCP setup to connect, and what’s available for the full tool list.
  • Sessions are now transcribed with Deepgram’s Nova 3 model for faster, more accurate transcripts.
  • A warning banner flags User Interviews participants who haven’t been paid within the required window.
  • Highlights and insights are now exposed as read endpoints, so an agent can pull the findings you’ve already captured, not just raw responses.
  • Fixed duplicate and orphaned meeting invites after moderator reassignment, calendars not refreshing after a reschedule, and Google Meet/livestream links stuck initializing.
  • Fixed reprocessing for recordings over 2GB, sessions missing from the repository after completion, and searchable-but-unlisted recordings.
  • Fixed MCP-created email templates rendering with an empty body.
​
In person studies
Participants can now provide an address for in-person interviews, so you can run studies that follow them into their own environment instead of a call.
  • Google address fields now return the full address after a selection is made, so the zip code is always included.
  • Observers can now view signed consent forms without seeing PII or downloading them.
  • The Members page now shows user email addresses, a role breakdown, name/email search, and filtering by role and status.
  • Eligibility rules now support “Any,” so a rule can match any filter instead of requiring all of them.
  • Cards can be set as default for billing and incentives.
  • Emails no longer get stuck or orphaned when a user’s connected account is removed.
  • Screener changes and edits persist, and library search works as expected.
  • Recordings now download as MP4 instead of FLV.
​
Shorter session timeout, for security
The “Remember me” login window was reduced from 30 days to 2 hours to meet security requirements.
  • Custom consent forms now show a preview instead of a thumbnail.
  • Sorting by participant rating always places unrated participants at the bottom, regardless of sort order.
  • Creators can request funding approval without the modal disappearing and incentives toggling off.
  • Studies funded by a Team wallet can now be closed when the refund exceeds the Team’s available funds.
  • Duplicating a study no longer errors if a former collaborator no longer has access.
​
Study Synthesis
A new AI Synthesis tab replaces the manual highlight board, auto-generating a structured summary once a study reaches 3 or more completed interviews, an overview, key insights, and major themes. Major themes carry inline citation badges that show supporting verbatim quotes in your session recordings, with chapter cards so you can scrub to individual moments.The legacy highlight board is still available under a “Highlights” sub-tab.
  • Unsaved-changes detection was added for eligibility rule edits.
  • Improved study duplication and sorting on the Repository table.
  • Corrected the URL used for the Databricks integration.
​

​
SCIM provisioning

EnterpriseEnterprise IT teams can now manage Great Question users directly from their identity provider through SCIM 2.0, automating provisioning, deprovisioning, and role management. Users are provisioned and deprovisioned automatically as your IdP changes, with name and email synced in real time. Push IdP groups to map to Great Question roles and Teams, with three sync modes (off, initial only, full sync). Configure it all at Settings → Governance → Provisioning, including bearer tokens, default roles, and IdP-managed roles and teams.

​
Data Retention v1.5

EnterpriseTake full control of your data retention policy with named rules for Recordings or Candidates, a preset or custom timeframe, and a deletion type. Rules run daily, with a validation step that shows exactly what will be deleted before activation and a 24-hour buffer before anything irreversible happens. Every rule creation, edit, status change, and deletion is recorded in a full audit trail, and rules can be paused or retired without losing that history.

​
Route approvals to specific admins

A new governance setting lets you route all approval requests to the account owner or a set of individual admins. When enabled, the approver dropdown is locked to that list across every request type, study publish, study funding, feature requests, and team requests.

​
Livestream status, visible everywhere

Livestream and observer bot status now shows up wherever you need it — the Interview room, participant table, slideout, and dashboard, so most issues no longer require a support ticket. You can resend the bot on demand from the Interview room, and pre-flight checks validate the URL, platform support, and meeting settings for Zoom, Meet, Teams, and Webex ahead of time, with proactive email notifications if something’s wrong.
  • Pagination now supports jumping to the first or last page from a clickable “Page X of Y” menu, across Candidates, Studies, Incentives, Repository, and in-study views.
  • A dedicated Ask AI tab was added to the Repository page.
  • Ad hoc messaging is now disabled for User Interviews participants, to prevent messaging conflicts.
  • Recordings now download with useful filenames (e.g. “Interview with Victoria.mp4”) instead of quality-based names.
​
Optional device permissions for unmoderated tests
Camera, microphone, and screen recording permissions can now be set to optional instead of required. When a permission is optional, participants are prompted to enable it but can skip and continue if they’d rather not, so you can collect richer data without blocking participants who can’t or won’t grant access. Available for all unmoderated studies; not available for AI-moderated studies.
  • Skip logic now works correctly in User Interviews studies.
  • Multi-select questions with “any of these” ideal answers now qualify correctly when a participant selects just one of them.
  • Auto-disqualification now works in User Interviews, matching how it behaves in Great Question.
​

​
User Interviews for external recruitment

We’ve partnered with User Interviews to bring you higher-quality participants: access to 6M+ verified participants (2x our previous panel size), a first match in under an hour, and industry-leading fraud detection, with 98.3% of sessions rated positive by research teams. You’ll continue to recruit through Great Question as usual, now with better quality, broader reach, and more consistent pricing.

​
AI Moderated Interviews

Conduct in-depth research at scale with AI that adapts in real time, probes for detail, and captures nuance, across time zones and hundreds of participants. Participants respond via video on their own schedule, and sessions flow directly into your repository. Available in English, French, German, Italian, Spanish, Russian, and Japanese. Interested in pricing? Reach out to sales@greatquestion.co.
  • Recruitment reach estimates and an estimated cost breakdown now appear for external recruitment requests.
  • Candidates can now be assigned an affiliation, a specific Team, the workspace, or no one, and the Study List has a new Affiliation column.
  • Snowflake now supports Key Pair authentication (Snowflake OAuth integrations are no longer supported).
Looking for something older? The full release history lives in the app. You can also check status.greatquestion.co for live service status.
Still need help? Contact us at support@greatquestion.co — median response time is 19 minutes during support hours.