{ "version": "https://jsonfeed.org/version/1.1", "title": "Zentrik Release Notes", "home_page_url": "https://zentrik.ai/releases", "feed_url": "https://zentrik.ai/releases.json", "description": "Recent shipped product updates from Zentrik. The complete history is available on the release archive.", "items": [ { "id": "https://zentrik.ai/releases/v1-51-0", "url": "https://zentrik.ai/releases/v1-51-0", "title": "v1.51.0: Find Signals by meaning and review trending evidence", "summary": "Find related Signals even when their wording differs from your query, preview the Evidence behind a trending Insight on the experimental Home Loop, and review Huddle asks by week. Clearer source checks and focused Evidence totals help teams make their next review.", "content_text": "Find Signals by meaning and review trending evidence\n\nv1.51.0 · Released 2026-10-15\n\nFind related Signals even when their wording differs from your query, preview the Evidence behind a trending Insight on the experimental Home Loop, and review Huddle asks by week. Clearer source checks and focused Evidence totals help teams make their next review.\n\nWhat changed\n\n- Search Signals and Insights by meaning alongside word matches. Signal results can appear when the wording differs from the query.\n- Preview the Evidence and Accounts behind recent Insight trends on the experimental Home Loop at /home/loop.\n- Open Huddle asks that need help, move between weeks, and see ask ages in calendar days.\n- Evidence topic totals count claims that match the topic. Existing Idea, Opportunity and Initiative links are visible on Evidence records again.\n- Invalid Signal source IDs now return a clear error before an import creates records or jobs. Put full URLs in sourceLinks.\n\nFind Signals by meaning\n\nSearch Discovery for Signals and Insights by words, IDs or meaning. A related Signal can appear even when it uses different words from your query, alongside matching Insights. Meaning search uses an embedding call and skips the meaning pass if it fails or takes more than three seconds. Account-name matching needs a query word of at least three letters.\n\n- Discovery search: /docs/product/customer-evidence-to-initiatives\n\nReview evidence behind a trend\n\nThe experimental Home Loop shows recent Signal trends. Preview a trending Insight to see its Evidence and contributing Accounts before you open it. The Home Loop is a separate preview at /home/loop; the existing Home page is unchanged. The preview uses the evidence available to your workspace.\n\nKeep Huddle asks in view\n\nHuddle puts blockers first in Need help, followed by overdue asks and then the oldest ask. Move between weeks with the week control, and use calendar-day ages to understand how long an ask has waited. Past weeks remain read-only; current-week actions depend on your team role.\n\nMore in this release\n\nFiltered Signal pages keep their filters, totals and pagination with less query-planning overhead. Invalid Signal source IDs are checked before an import creates records or jobs; keep IDs within 255 characters and put full URLs in sourceLinks. The pricing comparison shows SAML as an add-on for Team and Scale; plan entitlements do not change.\n\nMore in this release\n\nDiscovery\n\n- Meaning search can find Signals from their linked claims and mix them with Insights; word search remains available when meaning search is unavailable.\n- Evidence topic totals count matching claims, and links to Ideas, Opportunities and Initiatives appear on Evidence records.\n- Filtered Signal pages keep filters, totals and pagination. Invalid Signal source IDs fail before records or jobs are created.\n\nHome Loop and Huddle\n\n- The experimental Home Loop shows recent trends and previews evidence and Accounts behind an Insight.\n- Huddle Need help prioritizes blockers and overdue asks; the week control and calendar-day ages make timing clearer.\n\nNavigation and access\n\n- The navigation rail keeps destinations visible on laptop screens and places search beside the workspace name.\n- SAML is shown as an add-on for Team and Scale. Plan entitlements do not change.\n\nCanonical page: https://zentrik.ai/releases/v1-51-0\n", "date_published": "2026-10-15T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/semantic-signal-search/semantic-signal-search-light.png" }, { "id": "https://zentrik.ai/releases/v1-50-0", "url": "https://zentrik.ai/releases/v1-50-0", "title": "v1.50.0: Review Idea value and team asks", "summary": "Find the right Discovery record, inspect the estimated revenue that depends on an Idea, and see which teams are waiting on a decision. Tracked requests keep the people and customer context behind them.", "content_text": "Review Idea value and team asks\n\nv1.50.0 · Released 2026-10-15\n\nFind the right Discovery record, inspect the estimated revenue that depends on an Idea, and see which teams are waiting on a decision. Tracked requests keep the people and customer context behind them.\n\nWhat changed\n\n- Press ⌘K or Ctrl+K to find Discovery records by ID, words, spelling, Account or meaning, then open the best match.\n- Ask the assistant what recent tickets and calls are about: themes with counts, the newest customer quote and a link to each source.\n- Open an Idea’s Discovery → Value to review the estimated revenue range, the Accounts and deals behind it, and the effect of delaying it.\n- Review the Huddle week across teams, with owners, expected dates and overdue asks.\n- Preview tracked requests from Notion and Salesforce before bringing them into Ideas.\n\nFind an Idea and inspect its Value\n\nOpen an Idea’s Discovery → Value to review the estimated revenue range, the Accounts and deals behind it, and the effect of delaying it. The “This year” figure covers the next twelve months. The range depends on available customer evidence, commercial records and ship timing. It is an estimate for review, not promised revenue. Use the contributions and the question beneath them to check what the estimate rests on.\n\n- Idea Value and product terms: /docs/product/customer-evidence-to-initiatives\n\nAsk what customers keep raising\n\nAsk the assistant what recent support tickets and calls are about. It groups them by the insights already in your workspace, counts each theme and quotes the newest customer who raised it, with a link to the ticket or call. Ask a follow-up about the answer on screen, or ask it to trace an insight back to the signals, Accounts, opportunities and ideas behind it. Answers start sooner, and every record an answer names opens inside your workspace. The answer reflects the evidence your workspace holds; open the linked sources before you act on it.\n\nRead what teams are waiting on\n\nIn Huddle → All teams, open the teams that need help to see blockers, decisions, owners and expected dates together. Past weeks are read-only. For the current week, team editors and admins can assign or resolve asks. Workspace Owners and Admins can act when they belong to that team. Met counts recorded Huddle work; opening a Huddle alone does not count as a meeting.\n\nReview requests before importing\n\nBring Notion request tables or pages, and Salesforce product gaps, into Ideas with requester and Account context. Review the field mapping and existing-Idea matches before importing. In Discovery, choose Bring in evidence → Ideas you already track. The source must be accessible through the connected provider account. Notion reads the tables and pages shared with Zentrik; Salesforce product gaps keep their linked deal context.\n\n- Notion setup and imports: /docs/integrations/notion\n- Salesforce setup: /docs/integrations/salesforce\n\nKeep survey reporting within its source scope\n\nSurveyMonkey adds saved daily and historical imports, scoped NPS reporting and reviewed Account matching. Availability depends on provider access and approval. Select the connection, survey, question and Product for each report. Account matching stays reviewable, and anonymous or ambiguous responses are preserved. Account history shows the latest 50 receipts in the selected scope.\n\n- SurveyMonkey setup and limits: /docs/integrations/surveymonkey\n\nMore in this release\n\nConversations and voice\n\n- Resume your own saved assistant conversations from the header, with named record links in follow-ups.\n- Live transcription shows words while you speak again, and spoken or pasted feedback carries a clearer source label.\n\nConnections and review\n\n- Aha! imports retain comments and proxy-vote context, and can fill a missing customer requester without replacing one already set.\n- HubSpot setup is limited to workspace Owners and Admins. A revoked provider grant asks for reconnection without signing you out of Zentrik.\n- Review uncertain Product Feature suggestions under an Idea’s Related features. Your explicit choices, including clearing every link, take precedence over automatic matching.\n\nCanonical page: https://zentrik.ai/releases/v1-50-0\n", "date_published": "2026-10-15T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/quick-open-idea-value/quick-open-idea-value-light.png" }, { "id": "https://zentrik.ai/releases/v1-49-0", "url": "https://zentrik.ai/releases/v1-49-0", "title": "v1.49.0: Schedule campaigns and compare product bets", "summary": "Schedule reviewed email campaigns, weigh equally ranked Decision Lab bets, and inspect more complete product context and source details.", "content_text": "Schedule campaigns and compare product bets\n\nv1.49.0 · Released 2026-10-12\n\nSchedule reviewed email campaigns, weigh equally ranked Decision Lab bets, and inspect more complete product context and source details.\n\nWhat changed\n\n- Schedule a reviewed email campaign for a chosen date, time and one named time zone. Change or remove the schedule before dispatch.\n- Review equally ranked Decision Lab bets, their reasons and evidence, then choose a follow-up for yourself to review and send.\n- Inspect clearer source identity and quote context in Discovery, and use fuller scoped Context reads in Workspace Chat.\n\nSchedule a reviewed email campaign\n\nAdmin can schedule a reviewed email campaign batch for a chosen date, time and one named time zone. Change or remove the schedule before dispatch. A worker checks for due sends once a minute; actual inbox arrival can take longer.\n\nReview the Decision Lab slate\n\nWorkspace Chat shows Decision Lab results as equally ranked bets with readable reasons, linked ideas and accounts, and one confidence basis. Expand a bet to review its evidence and choose Review gaps, Draft opportunity or Drop. Close calls appear under Set aside. Follow-ups enter the composer for you to review and send; the interface does not commit a decision for you.\n\nInspect source and Context details\n\nDiscovery reviews can show a source's native rating and date, variant and provenance. A quote links to a contiguous original passage only when that source text is available; this does not verify generated wording. Workspace Context reads return full scoped bodies and mark any omissions. The assistant must fully read an omitted unit before it can call the context complete. Review cited sources because fuller retrieval does not prove an answer is correct.\n\nSet up workspaces and recover connections\n\nFirst-run guidance supports typed or spoken setup. Workspace setup is available through the API and CLI. Connected imports can recover a broken connection in place while keeping its known context. Viewer consent and saved-import key grants follow explicit scopes, and CLI reads respect server retry guidance.\n\nKeep Studies independent from Ideas Portal\n\nTeams can run Studies independently of an enabled Ideas Portal, with separate branding and friendly participant addresses. Existing links and workspace sign-in checks continue to work. This release did not test a customer identity provider.\n\nMore in this release\n\nHome and planning\n\n- The Home map remains optional; the native Home page stays the default.\n- Decision Lab keeps within its time budget and makes its ranked actions and comparison reasons easier to review.\n- Huddle and evidence models are updated. Members still review any proposed plan change before it is saved.\n\nAssistant and documents\n\n- Assistant turns record their outcomes, and answer typography stays with its answer.\n- Save a tailored document template without supplying a name or description.\n\nResearch and public labs\n\n- Compare text models in Public Labs while Responses stays off by default.\n- Research models and Study presentation have been updated. Review generated findings before relying on them.\n\nReliability\n\n- Onboarding and drafting models are updated, and stalled provider attempts can retry within the request time limit.\n- Connection and database recovery has improved. Deterministic job failures stop instead of repeating.\n\nCanonical page: https://zentrik.ai/releases/v1-49-0\n", "date_published": "2026-10-12T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/campaign-schedule-reviewed-batch/campaign-schedule-reviewed-batch-light.png" }, { "id": "https://zentrik.ai/releases/v1-48-0", "url": "https://zentrik.ai/releases/v1-48-0", "title": "v1.48.0: Search evidence and choose clearer next steps", "summary": "Search stored Evidence by meaning or exact words, inspect Assistant turn receipts, and choose a customer-backed idea before making it an Initiative.", "content_text": "Search evidence and choose clearer next steps\n\nv1.48.0 · Released 2026-10-11\n\nSearch stored Evidence by meaning or exact words, inspect Assistant turn receipts, and choose a customer-backed idea before making it an Initiative.\n\nWhat changed\n\n- Search stored Evidence by meaning or exact text, then inspect the records and results used in an Assistant turn.\n- Start an empty Initiatives list from an Idea or Insight backed by customer feedback, with a member choosing before commit.\n- Compare store ratings with an earlier completed Research run while keeping answers and themes on the immediately previous run.\n\nSearch Evidence and inspect Assistant turns\n\nSearch stored Evidence by meaning or exact words. The Assistant shows which record a turn used, lists record changes with their outcomes, and explains why it stopped. It qualifies incomplete retrieval. Write actions require your approval.\n\nChoose an Idea before making it an Initiative\n\nWhen a team has no Initiatives, the page can suggest an Idea or Insight with customer feedback behind it. A curated quote appears when available. A team member chooses before the item becomes an Initiative.\n\nCompare store ratings across Research runs\n\nResearch reports let you compare store ratings with any earlier completed run that saved ratings. Answers and themes continue to compare with the immediately previous run.\n\nEmail, imports, and reliability\n\nAdmin Email can save audiences and apply contact exclusions when you prepare product updates. Delivery rechecks the current rules and the audience criteria you reviewed. SurveyMonkey's existing response import is easier to find in the integration catalogue and Discovery. New connections remain limited until the provider approves the app; existing connected workspaces can continue importing. Invitation email budgets now follow the workspace plan. Results explain paused email counts and when an administrator can retry.\n\nReview Ideas and recover reads\n\nThe Idea source reader shows original material and identifies its source. Customer-facing wording needs review before publication. Roadmap reads retry one recognized transient connection failure. The app does not repeat writes. We patched security issues in the template and optional MCP client dependencies.\n\nMore in this release\n\nAccounts and Evidence\n\n- Account lists now handle large workspaces more efficiently while keeping filters, totals, ordering, and visibility rules.\n- Signal searches reuse current Account and Evidence relationships instead of recalculating them for every result.\n\nCanonical page: https://zentrik.ai/releases/v1-48-0\n", "date_published": "2026-10-11T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/initiative-first-commitment/initiative-first-commitment-light.png" }, { "id": "https://zentrik.ai/releases/v1-47-0", "url": "https://zentrik.ai/releases/v1-47-0", "title": "v1.47.0: Review weekly research and team work", "summary": "Initiative Review brings recorded work into one weekly view, while App Store research can group new reviews into counted themes.", "content_text": "Review weekly research and team work\n\nv1.47.0 · Released 2026-10-08\n\nInitiative Review brings recorded work into one weekly view, while App Store research can group new reviews into counted themes.\n\nWhat changed\n\n- Review weekly Initiative work by team, owner, or customer, then open the records behind each count.\n- Group five or more new App Store reviews into counted themes and inspect the source reviews.\n- Admins can review campaign content and eligible recipients before preparing tests and dispatch.\n\nReview the week in one place\n\nOpen Review from an Initiative to see its Studies, responses, Evidence, Delivery, decisions, and recorded needs. Group the work by team, owner, or customer, then open the records behind a count. Review starts with Your teams; owners and admins can widen it to the workspace. Huddle can show plans across teams. These views reflect recorded work.\n\nFind patterns in App Store reviews\n\nWhen a research run reads at least five new App Store reviews, it can group them into counted problems, requests, and praise. Compare themes with the previous run and open the source reviews behind a theme. The themes describe that sample; they do not represent overall customer sentiment.\n\nReview campaigns before dispatch\n\nAdmins can review uploaded email content and eligible recipients, freeze a batch, and prepare tests before dispatch. Dispatch is limited to recipients confirmed as queued. The release review did not dispatch a campaign or check inbox delivery.\n\nMore in this release\n\nHome and connected work\n\n- Home separates active delivery from completed work. Not now defers the current recommendation for seven days in this browser; a changed one can appear sooner.\n- Jira imports preserve code, inline code, and literal JSON. Refresh existing imported text for the correction. Evidence navigation respects reviewer-excluded relationships.\n\nCanonical page: https://zentrik.ai/releases/v1-47-0\n", "date_published": "2026-10-08T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/initiative-weekly-review/initiative-weekly-review-light.png" }, { "id": "https://zentrik.ai/releases/v1-46-0", "url": "https://zentrik.ai/releases/v1-46-0", "title": "v1.46.0: Keep trials and next steps on track", "summary": "Workspace admins can set a trial end date and owners receive reminders. Home keeps the Next step card clear while the trial guide is open.", "content_text": "Keep trials and next steps on track\n\nv1.46.0 · Released 2026-10-07\n\nWorkspace admins can set a trial end date and owners receive reminders. Home keeps the Next step card clear while the trial guide is open.\n\nWhat changed\n\n- Workspace owners and admins get email reminders three days before an eligible trial ends and after it lapses.\n- Admins can choose a trial end date from the Workspaces table, up to 120 days ahead.\n- Home keeps the Next step card beside Context and Discover, clear of the trial guide and other corner overlays.\n\nSet a trial end date\n\nFor eligible active trials, workspace owners and admins receive a reminder three days before the end date and another after the trial lapses. Admins can choose an end date from Admin → Workspaces, up to 120 days ahead; a later date already in place stays put. Each message includes workspace usage, Free plan limits, and a link to its Billing page.\n\nKeep the next step in reach\n\nHome places the Next step card at the foot of Context and Discover. It stays clear of the trial guide and other corner overlays while the remaining lanes keep their full height.\n\nLoad studies from external prototypes\n\nPrototypes hosted outside Zentrik can load their public study configuration and script. Other API routes keep the app's existing origin checks.\n\nMore in this release\n\nInterface polish\n\n- Press Escape to close an open dropdown inside a dialog.\n\nCanonical page: https://zentrik.ai/releases/v1-46-0\n", "date_published": "2026-10-07T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/trial-end-date-admin/trial-end-date-admin-light.png" }, { "id": "https://zentrik.ai/releases/v1-45-0", "url": "https://zentrik.ai/releases/v1-45-0", "title": "v1.45.0: Bring in evidence and choose the next step", "summary": "Discovery now keeps imports, evidence review, study planning and the next action closer together, with clearer workspace controls.", "content_text": "Bring in evidence and choose the next step\n\nv1.45.0 · Released 2026-10-07\n\nDiscovery now keeps imports, evidence review, study planning and the next action closer together, with clearer workspace controls.\n\nWhat changed\n\n- Bring files, notes and connected sources into Discovery through one guided import flow.\n- Review scoped and total evidence counts beside each Insight.\n- See one recommended next step on Home, based on current work and timing.\n- Keep Account membership current while preserving independent human links.\n\nBring evidence into one flow\n\nStart from what your team has: paste notes, add files or choose a connected source. Spreadsheet headers get editable suggestions, and an import receipt stays visible while background work continues. Export Excel files as CSV and choose unresolved Account names yourself. This flow does not merge Accounts or infer workbook hierarchy.\n\n- How Discovery uses imported data: /docs/product/import-evidence-files\n\nRead evidence in context\n\nInsight rows show both the evidence in your current scope and the total accepted links across dates and lifecycle states. Totals exclude evidence from other workspaces and items you manually excluded. Account membership also follows current source links while keeping independent human choices, votes and portal links intact. Reviewed source corrections and customer reconciliation remain separate work.\n\nChoose a useful next step\n\nHome offers one recommended move based on current Home data. It ranks deadlines, blocked initiatives, customer findings and work ready to advance by timing and consequence.\n\nReview and carry Study plans\n\nSelf-guided and AI-guided Study drafts now form in the layout used for review. Live interview guides keep their existing layout. For a ready builder brief, add a prototype Study and keep its instrumentation installed. Use an Owner or Admin action for Jira handoff; additional ticket and study links are best effort.\n\nMore in this release\n\nPlanning and workspace\n\n- Invitation controls show wait times and send failures; superusers can schedule one global service notice across the app and status page.\n- Billing separates Stripe-managed and Zentrik-managed plans, adds invoice access and prevents duplicate checkout. Stripe portal access still needs approved setup.\n- API and MCP clients can resolve public Idea references when creating or updating Insights; enterprise API rate profiles remain operator-managed.\n- Keep related initiative context and task names through planning updates, including work without key results or a PRD.\n\nConnections and reliability\n\n- Retry selected failed reads once without replaying writes; a persistent database or queue outage still needs recovery.\n- Reserve connection capacity for Insight synthesis and reduce unused generation and repeated tool output.\n- Send ideas and evidence to Jira while preserving edits to human-written descriptions.\n- Notify once when an automation stops because of its plan, and keep workbench filters and saved replay layouts in place.\n\nCanonical page: https://zentrik.ai/releases/v1-45-0\n", "date_published": "2026-10-07T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/signal-import-background-jobs/signal-import-background-jobs-light-1440x900.png" }, { "id": "https://zentrik.ai/releases/v1-44-2", "url": "https://zentrik.ai/releases/v1-44-2", "title": "v1.44.2: Guided research, clearer briefings and study audiences", "summary": "Plan public research, compare saved findings, coordinate Huddle follow-ups, and build Study audiences from existing People.", "content_text": "Guided research, clearer briefings and study audiences\n\nv1.44.2 · Released 2026-10-05\n\nPlan public research, compare saved findings, coordinate Huddle follow-ups, and build Study audiences from existing People.\n\nWhat changed\n\n- Review a guided research plan before starting, then inspect cited findings and compare saved answers without starting a paid research run.\n- Start Huddle from a team briefing with goals, customer evidence, and pending actions, while keeping saved layouts.\n- Build Study panels from existing People and include non-excluded respondents in Audience; filtered exports match the visible rows.\n- See when model-spend figures are partial estimates and how they relate to billed totals.\n\nPlan and compare public research\n\nChoose a guided Competition, Market, or Public feedback plan, review its scope, and start research when ready. Findings include cited sources and coverage limits. Compare saved subjects against shared questions using saved answers; comparison does not start a paid research run. App Store research uses a recent US storefront sample and is not a representative sentiment score. Agents and API clients can inspect or manage research with separate read, run, and manage permissions.\n\nStart with a Huddle briefing\n\nThe team briefing brings goals, customer evidence, and pending actions into one view. Follow links to the source items or act on overdue work. New users see three starter widgets; optional widgets remain available, and saved layouts stay in place.\n\nBuild Study audiences from People\n\nCreate a Study panel from existing People with search, filters, and selection across pages. Audience includes non-excluded respondents, and filtered exports match the visible rows. Panel creation saves selected members, not a live filter; creation supports up to 2,000 People per request. Response-only entries do not create People or invitations.\n\nRead model-spend estimates with their limits\n\nModel-spend reports now price proven short requests per request and label incomplete dollar figures as partial estimates. Billed totals remain visible; ambiguous tier mixes do not receive a dollar estimate. A partial estimate is not billed spend.\n\nReplay contextual guides while you work\n\nReplayable guides now explain screens across Initiatives, Discovery, Studies, Delivery, Objectives, Roadmap, Products, Context, Accounts, and People. Your progress stays in place while you complete an action and return to the guide. Existing role permissions and opt-outs still apply.\n\nMore in this release\n\nNavigation and conversations\n\n- Home keeps initiative scoping under Define and retains the compact next-step shortcut.\n- The Discovery assistant stays open while records refresh. Explicit record navigation keeps its usual handoff.\n- Chat record links use readable names and open the matching inspector or record page. Incomplete destinations remain readable text.\n\nEvidence and reliability\n\n- Evidence titles have less visual prominence. Transcript quote edits rebuild citations from spoken words; existing quotes are not updated automatically.\n- Regulated Brief emails omit source excerpts, but summaries, titles, charts, and links may still contain identifying information.\n- Homepage motion uses direct transitions with a still for reduced motion. Security pages clarify enforced product controls; selected dependencies were updated.\n\nCanonical page: https://zentrik.ai/releases/v1-44-2\n", "date_published": "2026-10-05T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/research-evidence-review/research-evidence-review-light.png" }, { "id": "https://zentrik.ai/releases/v1-44-1", "url": "https://zentrik.ai/releases/v1-44-1", "title": "v1.44.1: Review agent runs and compare prototypes", "summary": "Check an agent's saved instructions before a run, compare prototypes in Gallery or Table, and copy draft Study links before launch.", "content_text": "Review agent runs and compare prototypes\n\nv1.44.1 · Released 2026-10-05\n\nCheck an agent's saved instructions before a run, compare prototypes in Gallery or Table, and copy draft Study links before launch.\n\nWhat changed\n\n- Review the saved source, work unit, and launch setup before starting an agent run. Changes to reviewed inputs require a new review.\n- Switch Validation prototypes between Gallery and Table while filters, page size, and visible pagination stay in place.\n- Copy a participant link from a portal-based draft Study before launch. Copying does not publish the Study or send an invitation.\n- Use Jev and Hot or Not sample flows before live access is enabled; sample choices show recorded results without a provider request.\n\nReview an agent run before it starts\n\nFrom a saved Initiative source, review the saved instructions, selected work unit, its plan, deliverable summary, and launch setup before starting an execution agent. If a reviewed input changes, Zentrik asks you to review the new version first. Existing protection against duplicate or uncertain launches remains in place.\n\nCompare prototypes in Gallery and Table\n\nOpen Validation, then Prototypes, to compare linked prototypes as cards or rows. Filters and page size stay in the URL, and pagination remains visible while results scroll.\n\nCopy a draft Study link\n\nFor a portal-based draft Study, copy its participant link before launch. Copying does not publish the Study or send an invitation. The companion portal release 1.34.1 is deployed and checked first; private drafts remain unavailable to participants.\n\nTry public labs with a recorded sample\n\nJev and Hot or Not show useful sample flows when live analysis is off. Choosing a sample displays its recorded result; a photo preview stays in your browser. Live analysis still needs configured access and your explicit consent.\n\nMore in this release\n\nPlanning and reliability\n\n- Failed Persona edits keep the browser draft. Zentrik confirms only values the server saved before you leave the page.\n- Scheduled session refresh can recover from temporary failures with capped retries, then stops after invalidation or cleanup.\n- Invalid MCP tool inputs now return bounded correction guidance. Rejected calls still do not run, and authorization checks remain in place.\n\nProduct guidance and upkeep\n\n- The homepage carries one example from the first ticket through the next signal. Phone and reduced-motion layouts keep a complete static story.\n- Sample mode for the Jev and Hot or Not labs makes no live provider request; a live photo or call requires explicit consent when live access is available.\n- The release also updates selected security and browser dependencies and clarifies public integration and participant guides.\n\nCanonical page: https://zentrik.ai/releases/v1-44-1\n", "date_published": "2026-10-05T00:00:00Z", "image": "https://zentrik.ai/images/releases/v1.44.1/prototypes-table-light.png" }, { "id": "https://zentrik.ai/releases/v1-44-0", "url": "https://zentrik.ai/releases/v1-44-0", "title": "v1.44.0: A clearer Home and more control over imports", "summary": "Find the next action from Home, choose the size of your first import, and keep saved planning work intact as it changes.", "content_text": "A clearer Home and more control over imports\n\nv1.44.0 · Released 2026-10-04\n\nFind the next action from Home, choose the size of your first import, and keep saved planning work intact as it changes.\n\nWhat changed\n\n- Follow the work from discovery through delivery, with current-cycle progress on Home.\n- Set source scope, product routing and a smaller first batch before importing.\n- Find older Insights when the default window is empty, while keeping your filters.\n- Keep local task edits and reviewed execution agents through updates and retries.\n\nStart with the next useful action\n\nHome now groups work into Discover, Define, Validate and Deliver, with Context alongside them. Open Studies or Prototypes from Validate. Delivery shows the current cycle and its completed tasks; GitHub pull requests remain available as shortcuts.\n\nChoose what comes in first\n\nFirst import setup separates source scope, product routing and run controls. Automatic product routing is the default; choose a fixed product when you need one. For Zendesk, Zoom and Fathom, Per run lets you choose a smaller batch. This limit applies to one run, not the total history you can import. Detailed setup keeps your supported scope, routing, limit and cadence choices.\n\n- Set up Zendesk imports: /docs/integrations/zendesk\n\nLook beyond an empty month\n\nWhen the default 30-day Insights feed is empty, Zentrik checks 90 days and then 1 year, stopping when it finds results. Other filters stay in place, and a date range you choose takes precedence. A dismissible notice explains the expanded range; a failed request stops the search.\n\nKeep reviewed planning work\n\nManual initiative task batches now save together, and Jira or Linear refreshes preserve locally managed task fields. Story deletions from initiative chat appear in activity history. Save an edited execution plan before generating agents. Replacing saved agents requires confirmation, and a launch with an unknown outcome stays on hold for review.\n\nMore in this release\n\nNavigation and review\n\n- Opportunities opens in Tree. Choose Tree, List or Themes from the toolbar; existing links to a specific view keep their meaning.\n- Knowledge navigation now lists Products, Accounts, Objectives and Context. The collapsed Products picker includes All products.\n- Idea readiness uses Evidence support, Problem clarity and Solution clarity; the underlying scores are unchanged.\n\nEvidence and connections\n\n- Narratives use explicitly accepted evidence, and Knowledge checks quotes against source text. You still need to review whether the evidence supports the conclusion.\n- MCP connections handle verified Claude host metadata and identify replacement actions more clearly. Existing roles and consent still apply.\n- Jira setup checks project access before offering an import; reconnect when access is rejected.\n\nCanonical page: https://zentrik.ai/releases/v1-44-0\n", "date_published": "2026-10-04T00:00:00Z", "image": "https://zentrik.ai/images/releases/v1.44.0/home-delivery-observed-light.png" }, { "id": "https://zentrik.ai/releases/v1-43-0", "url": "https://zentrik.ai/releases/v1-43-0", "title": "v1.43.0: Clearer discovery and safer planning", "summary": "Carry record context into discovery chat, review bounded import progress, and approve objective measures only after you review them.", "content_text": "Clearer discovery and safer planning\n\nv1.43.0 · Released 2026-10-04\n\nCarry record context into discovery chat, review bounded import progress, and approve objective measures only after you review them.\n\nWhat changed\n\n- Open discovery chat with the current record in context and follow customer signals toward ideas.\n- See Jira scope and progress, and run bounded Pendo history imports with recovery controls.\n- Review objective and measure drafts before a permission-checked acceptance saves them.\n\nKeep discovery work in context\n\nOpen chat from a discovery record with that record in context, then follow Signal learnings toward related ideas. Study authors can set interview length and organize scripts around what participants say. The chat view does not judge answer quality for you.\n\n- Discovery context and Signals: /docs/product/discovery-context-and-signals\n- Explore Studies: /docs/product/idea-studies\n\nReview connected import progress\n\nAfter you choose Jira scope, the import view shows progress and the number of records in that scope; selecting scope does not start the import. Pendo history runs use three-day windows and a shared limit of 25 new, retried, or processing Signals. Gong retries use a 24-hour overlap for late transcripts; an older failed call needs an explicit retry.\n\n- Jira integration guide: /docs/integrations/jira\n- Pendo integration guide: /docs/integrations/pendo\n- Gong integration guide: /docs/integrations/gong\n\nReview objective measures before saving\n\nDraft objectives and suggested measures stay separate from saved goals. A workspace editor reviews and confirms a proposal before Zentrik accepts it; the permission check and retry guard protect the save. This synthetic example shows a proposal waiting for review, not a live model call or a saved measure.\n\nMore in this release\n\nIdeas and collaboration\n\n- Shared Slack and record links open in their named workspace with an unavailable-record explanation.\n- Help separates AI support conversations from explicit requests to the team.\n\nPlanning and workspace\n\n- Human edits and linked work stay in place through story generation and retries.\n- Undo appears only when the workspace can restore the removed Initiative.\n- The ChatGPT catalog asks for explicit scope consent; publication and installed-client checks remain separate.\n\nCanonical page: https://zentrik.ai/releases/v1-43-0\n", "date_published": "2026-10-04T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/objective-measure-review/objective-measure-review-light.png" }, { "id": "https://zentrik.ai/releases/v1-42-0", "url": "https://zentrik.ai/releases/v1-42-0", "title": "v1.42.0: Study drafts and decision briefs", "summary": "Watch an interview guide take shape, use authorized Zentrik tools from modern MCP clients, and review read-only decision briefs in ChatGPT.", "content_text": "Study drafts and decision briefs\n\nv1.42.0 · Released 2026-10-03\n\nWatch an interview guide take shape, use authorized Zentrik tools from modern MCP clients, and review read-only decision briefs in ChatGPT.\n\nWhat changed\n\n- Watch an AI-guided interview draft form, then edit the completed guide in the same dialog.\n- Connect modern and legacy MCP clients to authorized Zentrik tools with stricter input checks.\n- Review a read-only recommendation in ChatGPT and choose a follow-up that continues the conversation.\n\nReview an interview guide as it forms\n\nIn the Study dialog, an AI-guided interview draft streams into view. When it is ready, review and edit the guide in that same dialog. This preview uses synthetic data; it does not show a live model response or create a Study.\n\n- Explore Studies: /docs/product/idea-studies\n\nUse Zentrik tools from MCP clients\n\nModern MCP clients can call the Zentrik tools authorized for their workspace. Legacy clients continue to work, and malformed business inputs are rejected before they reach the data store. Host model metadata can pass through without relaxing those input checks.\n\n- MCP overview: /docs/integrations/mcp\n\nReview a decision brief in ChatGPT\n\nChatGPT can show a read-only recommendation and offer two follow-up choices that continue the conversation. Tool approval, the public host check, and installed-client checks still determine whether a person can use it in their ChatGPT setup.\n\n- ChatGPT integration guide: /docs/integrations/mcp-chatgpt\n\nMore in this release\n\nDiscovery and planning\n\n- Study authors can edit a completed interview guide in the dialog where they drafted it.\n- Decision briefs in ChatGPT are read-only; follow-up choices continue the conversation.\n\nConnected tools\n\n- Modern and legacy MCP clients use workspace-authorized tools, with stricter business-input validation.\n- ChatGPT availability still depends on host approval and installed-client checks.\n\nCanonical page: https://zentrik.ai/releases/v1-42-0\n", "date_published": "2026-10-03T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/live-interview-draft-stream/live-interview-draft-stream-light.png" }, { "id": "https://zentrik.ai/releases/v1-41-0", "url": "https://zentrik.ai/releases/v1-41-0", "title": "v1.41.0: More control over Studies and initiative capacity", "summary": "Build study audiences, choose a method, and opt into shared question flows. Initiative capacity now follows status, with add-ons available in the app.", "content_text": "More control over Studies and initiative capacity\n\nv1.41.0 · Released 2026-10-03\n\nBuild study audiences, choose a method, and opt into shared question flows. Initiative capacity now follows status, with add-ons available in the app.\n\nWhat changed\n\n- Choose a study audience from idea contributors, a team panel, or a pasted list.\n- Start a study with one method chooser and opt into a shared, answer-aware question flow.\n- See initiative capacity by status, keep Scale limits during setup, and buy more capacity in app.\n- Move from a call's evidence to its ideas without losing the call context.\n\nChoose who a study is for\n\nBuild an audience from people who asked for the idea, a team panel, or a pasted list. Zentrik does not send invitations; your team copies emails or downloads a CSV and invites people through its own channel. A recent answer counts as prior study input for 14 days. Validation studies hold back people who already gave input on their ideas.\n\n- Study audiences guide: /docs/product/study-audiences\n\nChoose a method and shape the questions\n\nOne chooser offers self-guided, prototype, interview, and AI-guided methods. Optional shared flows carry short answers into surveys, AI interview instructions, and live guides. A short answer can turn “login flows” into “For login flows, what would help most?” Declined or long answers leave the next question plain. Participant surveys use Portal 1.33.0; the AI interview screen stays the same.\n\n- Explore Studies: /docs/product/idea-studies\n- Test a prototype: /docs/product/test-prototype\n\nPlan and expand initiative capacity\n\nThe initiative limit now counts work by status. Scale limits stay available during the setup window. Paid workspaces can buy initiative capacity in the app, and teams can archive initiatives to make room.\n\nFollow customer evidence into ideas\n\nWhen a call has qualifying ideas, its Ideas tab pairs each idea with the evidence excerpt from that call. Open the excerpt, Signal, or idea from the same view. Members with Discovery write access can rename Signals from the inspector or record page. Interview Signals open directly, with a status indicator while processing.\n\nMore in this release\n\nIdeas and support\n\n- Full-page idea overviews show attachments with image and text previews, downloads, and uploads for writers.\n- Support ticket replies can include one PNG, JPEG, WebP, or GIF image up to 10 MB without accompanying text.\n- Idea creation validates linked records within the workspace and saves the idea and its selected links together.\n\nResearch and other updates\n\n- Web research now runs on GPT-6.1 Sol. Report and notes extraction can recover when a model reply exceeds its output limit.\n- Community discussion grounding and batch correction of contradictory positive feedback are more reliable.\n- Evidence backfill can process a chosen batch of 1 to 1,000 Signals, newest first.\n- Study titles and internal stages can be edited after launch while participant-facing questions and settings stay locked.\n- Gong film pages and launch clips have clearer, consistent viewing controls.\n\nCanonical page: https://zentrik.ai/releases/v1-41-0\n", "date_published": "2026-10-03T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/study-method-chooser/study-method-chooser-light.png" }, { "id": "https://zentrik.ai/releases/v1-40-0", "url": "https://zentrik.ai/releases/v1-40-0", "title": "v1.40.0: Richer imports and clearer initiative routing", "summary": "Aha! imports carry saved mappings and source context. Insight review includes customer workflows and unmet needs, while Initiative Overview can route open questions. New public pages bring Zentrik films together.", "content_text": "Richer imports and clearer initiative routing\n\nv1.40.0 · Released 2026-10-01\n\nAha! imports carry saved mappings and source context. Insight review includes customer workflows and unmet needs, while Initiative Overview can route open questions. New public pages bring Zentrik films together.\n\nWhat changed\n\n- Search through up to 5,000 Aha! products and reuse a saved mapping in scheduled imports.\n- Capture and classify Customer workflow and Unmet need Insights across the UI and API.\n- Assign an open question or start an Initiative brief while setup questions remain.\n- Browse Zentrik films on shareable pages and choose available wide or vertical cuts.\n\nCarry more context through Aha! imports\n\nProduct search can page through up to 5,000 products. A manual import remembers its full mapping for later scheduled runs. Re-imports follow Aha! status until someone changes the Zentrik idea, and imported source context is retained. Unreadable custom-field values are skipped and reported. After organizations are imported, teams can link idea voters to accounts. Comments, numeric custom-field types, proxy-vote values, and voting organizations are not included in these outcomes.\n\nClassify customer workflows and unmet needs\n\nCustomer workflow and Unmet need are now available in Insight capture and classification across the UI and API. Existing Insights in Other are not reclassified automatically; a team can recategorize an Insight after deployment.\n\nRoute open questions from Initiative Overview\n\nFrom Initiative Overview, assign an open question to someone and see who the team is waiting on. Editors can start a brief while questions remain open. Assigning a question uses the same write-access rule as answering it.\n\nBrowse the Zentrik films library\n\nZentrik’s films library brings the films together on shareable pages. Viewers can choose wide or vertical cuts where both are available and copy a film link. The Gong page explains the read-only call-to-decision workflow and links to the Gong Collective listing; its film uses illustrative sample data.\n\n- Browse Zentrik films: /loops/films\n- Gong integration page: /features/gong-integration\n\nMore in this release\n\nJira import guidance\n\n- Daily Jira import guidance now names Team as the first eligible plan. The entitlement check itself is unchanged.\n\nCanonical page: https://zentrik.ai/releases/v1-40-0\n", "date_published": "2026-10-01T00:00:00Z", "image": "https://zentrik.ai/images/product-narratives/gong-call-to-decision/gong-call-to-decision-portrait-poster.jpg" }, { "id": "https://zentrik.ai/releases/v1-39-0", "url": "https://zentrik.ai/releases/v1-39-0", "title": "v1.39.0: Guided trials and clearer Jira previews", "summary": "Trial workspaces can follow a guided route from connected signals to a first idea. Jira previews show issue types and frequent reporters, and Signals keeps an account filter in scope as you move through results.", "content_text": "Guided trials and clearer Jira previews\n\nv1.39.0 · Released 2026-09-30\n\nTrial workspaces can follow a guided route from connected signals to a first idea. Jira previews show issue types and frequent reporters, and Signals keeps an account filter in scope as you move through results.\n\nWhat changed\n\n- Follow a guided route from a connected signal source to your first idea in an eligible trial workspace.\n- Review Jira issue types, frequent reporters, generated JQL, and the issue count before saving an import.\n- Keep a selected account in Signals while you paginate and move through browser history.\n\nReach a first idea with a guided trial route\n\nThe tour guides eligible trial workspaces from a connected signal source through import and evidence review to creating or linking an idea. The tour appears only when the workspace plan is trialing; existing demo workspaces remain eligible. If source eligibility cannot be confirmed, Zentrik hides the invitation. The tour completes only after the idea is created or fully linked.\n\n- Explore the Jev Signal Map sample: /labs/jev\n\nReview Jira reporters before saving\n\nJira import previews show matching issue types and frequent reporters, so an owner can exclude a monitoring reporter and review the generated JQL and issue count before saving. Facets are sampled above the preview cap. Reporter names can be shared, so inspect the JQL and count before you save.\n\n- Jira imports guide: /docs/integrations/jira\n\nKeep the Signals account filter in scope\n\nThe account filter now narrows the Signals list and keeps the selected account in scope through pagination and navigation. It uses the existing API and account-matching rules; it does not change which Signals are linked to an account.\n\nMore in this release\n\nSignals filter\n\n- The selected Signals account appears as a removable label, so a person can clear the filter from the results view.\n\nCanonical page: https://zentrik.ai/releases/v1-39-0\n", "date_published": "2026-09-30T00:00:00Z" }, { "id": "https://zentrik.ai/releases/v1-38-0", "url": "https://zentrik.ai/releases/v1-38-0", "title": "v1.38.0: Focused Jira imports and cited research answers", "summary": "Jira setup now shows the source, context, and first batch before a draft runs. Discovery Brief can show saved research answers with citations, and Jev offers an editable public sample.", "content_text": "Focused Jira imports and cited research answers\n\nv1.38.0 · Released 2026-09-29\n\nJira setup now shows the source, context, and first batch before a draft runs. Discovery Brief can show saved research answers with citations, and Jev offers an editable public sample.\n\nWhat changed\n\n- Review Jira source, context, and the first batch before you run a draft import.\n- Read saved Web Research answers beside their cited passages in Discovery Brief.\n- Edit an illustrative sample in the public Jev Signal Map before choosing live analysis.\n\nReview a Jira import before the first run\n\nOpen Integrations → Jira to configure the source and context on a focused page. Draft runs default to 25 issues; you can choose 100, 500, or all, and add a creation-date boundary. The limit applies to draft runs. Ongoing imports keep their full saved source and cursor. A changed date boundary needs a fresh source check before you save.\n\n- Jira imports guide: /docs/integrations/jira\n\nRead saved research answers in Discovery Brief\n\nA Brief can show an already saved Web Research answer beside its cited passages and coverage gaps while the saved research definition and in-scope source still match. If an answer is missing or stale, the existing Evidence path stays available. Email detail follows each recipient’s current preference; showing the answer does not run a second research call.\n\nExplore the public Jev Signal Map\n\nThe public Jev Signal Map opens with an editable illustrative sample. Live analysis is optional and appears only when service configuration and your consent are in place; live runs share limits with the other public demos. The sample shows the review flow, not live model quality.\n\n- Open the Jev Signal Map sample: /labs/jev\n\nMore in this release\n\nDiscovery and workspace\n\n- Reviewed Learn items leave the default feed until new evidence arrives; Include reviewed and Show all can bring them back.\n- Study launch errors list setup blockers, and pending prototype edits save before launch.\n- Workspace switching keeps a valid section destination, and Home shows its time-saved breakdown again.\n\nImports and reliability\n\n- Account-only CRM runs stay within the reviewed batch; unresolved source parents keep the existing hierarchy until a source binding is confirmed.\n- Pendo imports retain source ownership for Account links; historical links with unclear ownership still need review.\n- Public responses use consistent security headers, safer redirects, bounded static-file serving, and verified upstream TLS.\n\nCanonical page: https://zentrik.ai/releases/v1-38-0\n", "date_published": "2026-09-29T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/saved-import-lifecycle/saved-import-lifecycle-light.png" }, { "id": "https://zentrik.ai/releases/v1-37-0", "url": "https://zentrik.ai/releases/v1-37-0", "title": "v1.37.0: Question-first research with cited Discovery findings", "summary": "Start research with a question, review a plan, and inspect cited findings before choosing what to add to Discovery.", "content_text": "Question-first research with cited Discovery findings\n\nv1.37.0 · Released 2026-09-28\n\nStart research with a question, review a plan, and inspect cited findings before choosing what to add to Discovery.\n\nWhat changed\n\n- Review an editable research plan before you start a bounded public-source run.\n- Inspect cited observations and coverage gaps, then choose the exact finding to add.\n- Search connected sources and review import status in one workbench.\n- Correct Signal speaker sides and dates, then share a Signal as an image.\n\nAsk a question, then review the evidence\n\nStart with a research question and review or edit the plan before you run it. Inspect cited observations and coverage gaps, compare runs, and choose whether to add an exact finding to a product. Public-source research shows a bounded sample; it does not establish complete coverage. Review citations and corrections before you use a finding.\n\nReview connected sources\n\nSources is a searchable workbench for supported saved imports. Review available details and received counts, then use the import history to see what arrived. New workspaces get guidance for connecting a first source. Provider answers in the onboarding preview are simulated.\n\n- Connected sources: /docs/integrations\n\nKeep Discovery actions consistent\n\nUse the same record actions across Discovery tables, inspectors, and pages. Evidence traceability makes inherited support easier to inspect. Review records and permissions before you apply changes.\n\nCorrect and share Signals\n\nPeople with write access can correct Signal dates and speaker sides. Roster details stay hidden by default and session roles reset automatically. Export a Signal as a Learnings, Overview, or Square image; these cards omit IDs, relevance, account names, and ARR.\n\nMore in this release\n\nDelivery and reliability\n\n- The Delivery timeline stays stable while you zoom, open initiatives, and edit dates.\n- Account product drafts remain during autosave.\n- MCP errors and analytics labels are clearer.\n\nCanonical page: https://zentrik.ai/releases/v1-37-0\n", "date_published": "2026-09-28T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/research-evidence-review/research-evidence-review-light.png" }, { "id": "https://zentrik.ai/releases/v1-36-0", "url": "https://zentrik.ai/releases/v1-36-0", "title": "v1.36.0: Clearer discovery and delivery, safer Jira and Zoom workflows", "summary": "Review source-backed Discovery summaries, keep Jira cycle history in view, and manage connected imports with clearer ownership and controls.", "content_text": "Clearer discovery and delivery, safer Jira and Zoom workflows\n\nv1.36.0 · Released 2026-09-27\n\nReview source-backed Discovery summaries, keep Jira cycle history in view, and manage connected imports with clearer ownership and controls.\n\nWhat changed\n\n- Review a saved Discovery Brief summary with its evidence links and source context.\n- Follow Jira sprint history alongside Zentrik delivery work.\n- Understand which imports are managed by a connection, and what Zoom disconnect removes.\n- Open record history from customizable properties.\n\nReview Discovery evidence\n\nA saved Brief can summarize recent Evidence with source names, dates, and links when your permissions allow. Treat the summary as a starting point, not a verified conclusion. Evidence links connect Ideas to their supporting records.\n\n- Discovery overview: /docs/product\n\nFollow work across Jira and Zentrik\n\nA Jira Scrum board maps to one Zentrik team. Jira cycle history updates every 15 minutes and shows the active sprint, up to six closed sprints, and three upcoming sprints. Jira remains the owner of synced records. Review Jira history when you need older change details; long changelogs can limit how far back Zentrik can show them.\n\n- Jira integration: /docs/integrations/jira\n\nManage connected imports\n\nSources now shows which imports are managed by a connection. When you disconnect Zoom, Zentrik deletes stored Zoom transcript content and speaker identity. Derived Insights, Evidence records, and their links remain. Some older imports may not have a disconnect timestamp, and provider or model logs and backups follow their own retention periods.\n\n- Zoom integration: /docs/integrations/zoom\n\nOpen record history from properties\n\nCustomizable record properties can include history, so you can review changes without leaving the record.\n\nMore in this release\n\nDiscovery and delivery\n\n- Discovery keeps context visible while summaries and evidence links load.\n- Workspace writes stay within the selected workspace and role.\n- Delivery views remain stable while you open initiatives and edit dates.\n\nCanonical page: https://zentrik.ai/releases/v1-36-0\n", "date_published": "2026-09-27T00:00:00Z" }, { "id": "https://zentrik.ai/releases/v1-35-0", "url": "https://zentrik.ai/releases/v1-35-0", "title": "v1.35.0: Clearer Discovery, planning, and source imports", "summary": "Find Ideas by meaning, review suggested changes before applying them, and plan initiative work with clearer ownership and next steps. Documents and meeting imports also show more of their progress where teams work.", "content_text": "Clearer Discovery, planning, and source imports\n\nv1.35.0 · Released 2026-09-27\n\nFind Ideas by meaning, review suggested changes before applying them, and plan initiative work with clearer ownership and next steps. Documents and meeting imports also show more of their progress where teams work.\n\nWhat changed\n\n- Find Ideas by meaning in Discovery alongside exact-text matches.\n- Reviewers can edit suggested document changes, user stories, and acceptance criteria before applying them.\n- See task roles, dependencies, and next steps on the Initiative Work map.\n- Choose and review meeting imports, including Zoom meetings across up to 31 days.\n\nSearch and review Discovery evidence\n\nDiscovery search now includes Ideas matched by meaning alongside the keyword path for exact wording and public IDs. Reviewers can edit suggested document changes, user stories, and acceptance criteria before applying them. Older Evidence Findings reports need a refresh before quick actions are available.\n\nPlan work by role and cycle\n\nThe Work map places tasks in role lanes and dependency order, then marks what can start and what is waiting. People with task-edit access can update task roles, links, and status on the map. Initiative Schedule replaces Milestones and Capacity as the plan timeline and uses task cycles. Existing milestone records and their links to tasks and key results have been removed.\n\n- Planning and roadmaps: /docs/product/planning-and-roadmaps\n- Schedule work on the roadmap: /docs/product/schedule-roadmap\n\nWrite documents where they will live\n\nAI-generated briefs, requirements, technical designs, and custom documents show section progress on their final page. The page reconnects to the writing progress after a reload and keeps the finished text visible while the saved version loads. Review generated text before using it.\n\n- Documents and delivery: /docs/product/documents-and-delivery\n\nReview meeting imports\n\nSources can continue from a connection into its first import. Zoom imports can review and select meetings across a date range of up to 31 days. Anyone with import access can add a personal Granola key; personal notes stay manual-review only and are not used by scheduled imports.\n\nMore in this release\n\nDiscovery Brief\n\n- Daily and weekly Briefs compare rolling seven-day periods, and a person's first email after setup arrives even when no new finding is ready.\n- An invitation can offer the recipient a link to start their own Brief. The recipient must save their choice; the workspace does not opt them in.\n\nSource access\n\n- A person with import access can add their own Granola key. Personal notes stay manual-review only and are not used by scheduled imports.\n\nCanonical page: https://zentrik.ai/releases/v1-35-0\n", "date_published": "2026-09-27T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/initiative-findings-preview-before-apply/initiative-findings-preview-before-apply-light.png" }, { "id": "https://zentrik.ai/releases/v1-34-0", "url": "https://zentrik.ai/releases/v1-34-0", "title": "v1.34.0: Review Evidence and shape initiative decisions", "summary": "Review the source and relevance of Discovery evidence, preview Initiative edits before you apply them, and ask questions that help a team make a decision.", "content_text": "Review Evidence and shape initiative decisions\n\nv1.34.0 · Released 2026-09-25\n\nReview the source and relevance of Discovery evidence, preview Initiative edits before you apply them, and ask questions that help a team make a decision.\n\nWhat changed\n\n- See clearer source labels and relevance details in Discovery.\n- Preview suggested Initiative edits before you apply them.\n- Review questions that name their decision and link to supporting context.\n\nReview evidence before you act\n\nDiscovery uses labels that name the source type. In Initiative Evidence Findings, you can open the source and preview a suggested document or user-story edit before you apply it. Older reports need a refresh before quick actions are available. Relevance decisions ask for confirmation; optional relevance memory only affects evidence for the same product and category.\n\nAsk questions that support a decision\n\nWhen Zentrik generates Initiative or Idea questions, each question explains the decision it supports. Open What Zentrik knows to review linked facts and sources, then use an answer option or edit a sourced draft. Sparse records can leave open decisions unanswered, so review each question and its sources before acting.\n\nMore in this release\n\nResearch and reliability\n\n- See the first Insights page and current counts together while you review a filtered feed.\n- For some long transcripts, extraction can retry in smaller sections after the first response is too long. Very large or canceled requests can still fail.\n\nCanonical page: https://zentrik.ai/releases/v1-34-0\n", "date_published": "2026-09-25T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/initiative-findings-preview-before-apply/initiative-findings-preview-before-apply-light.png" }, { "id": "https://zentrik.ai/releases/v1-33-0", "url": "https://zentrik.ai/releases/v1-33-0", "title": "v1.33.0: Manual Evidence capture and steadier workflows", "summary": "Add manually reviewed Evidence to Signals through the public API, and keep Initiative work areas and Accounts review usable as teams move through their work.", "content_text": "Manual Evidence capture and steadier workflows\n\nv1.33.0 · Released 2026-09-24\n\nAdd manually reviewed Evidence to Signals through the public API, and keep Initiative work areas and Accounts review usable as teams move through their work.\n\nWhat changed\n\n- Add a manually reviewed Evidence claim to a Signal through the public API\n- Keep Initiative work areas reachable when the workspace header has limited width\n- Keep selected Accounts columns while sorting, and use the displayed weighted score for Discovery sorting\n\nAdd reviewed Evidence to Signals\n\nAPI clients with Signal read and Evidence curation access can add a manually reviewed Evidence claim to a Signal. Optional extracts must match the stored Signal text. Each request creates a new Evidence item; after a timeout, check the Signal's Evidence before retrying.\n\n- Read the Signals API guide: https://zentrik.ai/docs/api/signals\n\nReach Initiative work areas in a narrow header\n\nKeep Initiative work areas reachable when the workspace header has limited width. Use the existing compact picker to reach another work area.\n\nKeep Accounts sorting consistent\n\nKeep selected Accounts columns while sorting, and use the displayed weighted score for Discovery sorting.\n\nMore in this release\n\nAPI guidance\n\n- Clarify MCP guidance for existing validation and concurrency limits, and pass both returned workflow handles when polling.\n\nCanonical page: https://zentrik.ai/releases/v1-33-0\n", "date_published": "2026-09-24T00:00:00Z" }, { "id": "https://zentrik.ai/releases/v1-32-0", "url": "https://zentrik.ai/releases/v1-32-0", "title": "v1.32.0: Follow evidence into people and planning", "summary": "Review source-backed Discovery patterns, carry customer context into shared records, and keep Initiative and Study work moving.", "content_text": "Follow evidence into people and planning\n\nv1.32.0 · Released 2026-09-24\n\nReview source-backed Discovery patterns, carry customer context into shared records, and keep Initiative and Study work moving.\n\nWhat changed\n\n- Trace market patterns to their supporting evidence and sources.\n- Sort, filter, and manage people across a workspace.\n- Keep Initiative, Study, and Discovery Brief work connected.\n\nReview Discovery evidence by source\n\nOpen a market pattern to inspect its supporting Signals and Evidence. Market Insights use independent, unscoped Signals; product-specific evidence remains scoped to its product. Source lineage and Evidence extracts are easier to audit, while sparse and dense views keep their own useful detail.\n\nCarry customer context into shared people records\n\nUse a workspace-wide People directory to see each person's role, Account, conversations, Studies, and votes together. Owners and admins can sort columns, filter by role, source, Studies, and votes, then assign or delete selected people. Completed conversation imports show their Signals, Evidence, Insights, and People receipt together.\n\nPlan Initiatives and Studies with clearer context\n\nWorkspace-safe Initiative links and edit permissions keep work in the right scope. Study follow-up questions, participant identity, and response handling stay connected through setup and participation. These changes do not claim a research finding or participant outcome.\n\nKeep Discovery Brief delivery useful\n\nExpand the Brief's evidence scope and save safe scheduled snapshots. Quiet weeks remain quiet, and email delivery skips an edition when it has no useful content.\n\nMore in this release\n\nConnected workflows\n\n- Improve account reconciliation and Open CRM entry points.\n- Keep workspace-chat tool results with the final answer and preserve partial text after early interruption.\n- Keep Study links, invitation identity, and follow-up handling consistent.\n\nReliability and access\n\n- Use supported public API links and clearer privacy details.\n- Protect Initiative edits with workspace and assigned-team access checks.\n\nCanonical page: https://zentrik.ai/releases/v1-32-0\n", "date_published": "2026-09-24T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/trend-to-evidence-review/trend-to-evidence-review-light-poster.webp" }, { "id": "https://zentrik.ai/releases/v1-31-0", "url": "https://zentrik.ai/releases/v1-31-0", "title": "v1.31.0: Keep research sources and study work connected", "summary": "Carry Web Research citations, prototype context, and support ownership into the next step, with recovery and access checks for interrupted or restricted work.", "content_text": "Keep research sources and study work connected\n\nv1.31.0 · Released 2026-09-23\n\nCarry Web Research citations, prototype context, and support ownership into the next step, with recovery and access checks for interrupted or restricted work.\n\nWhat changed\n\n- Keep source citations when opening or searching within a Web Research source\n- Start a prototype task study with its selected revision and linked work\n- Route support alerts to request owners and filter the inbox by assignment\n- Limit Initiative document edits to users with the required workspace and team access\n\nCarry Web Research sources into review\n\nKeep source citations with Web Research findings when you open or search within a source. The research evaluator adds synthetic and public-evidence comparisons; production models and prompts remain unchanged, and small evaluations do not establish general quality or cost improvement.\n\nStart the right study from a prototype\n\nStart a prototype task study from the prototype, with its selected revision and linked work carried into configuration. Self-guided feedback and AI-guided interviews remain separate methods; this does not claim that a study has run or been validated with participants.\n\nKeep support requests with their owners\n\nRoute support alerts to an assigned owner's account email and filter the inbox to requests assigned to you. Unassigned requests keep the existing team recipients.\n\nProtect Initiative document edits\n\nRequire workspace and assigned-team access before editing Initiative document sections. The edit check does not change document read access.\n\nMore in this release\n\nConnected research\n\n- Stop saved research schedules after disconnect and preserve earlier acquisition evidence\n- Keep import evidence when a database connection interrupts a read\n\nStudies and collaboration\n\n- Keep added interview transcripts available while a Live Interview study is in draft\n- Align support email identity and delivery details across notifications\n\nAccess and reliability\n\n- Retry supported reads once after a database disconnect while surfacing persistent failures\n- Keep support email identity and delivery metadata consistent\n\nCanonical page: https://zentrik.ai/releases/v1-31-0\n", "date_published": "2026-09-23T00:00:00Z" }, { "id": "https://zentrik.ai/releases/v1-30-0", "url": "https://zentrik.ai/releases/v1-30-0", "title": "v1.30.0: Clearer first steps for Discovery and research", "summary": "Find a useful next step in onboarding, Discovery Briefs, and prototype studies, with clearer limits on what public research and study evidence can show.", "content_text": "Clearer first steps for Discovery and research\n\nv1.30.0 · Released 2026-09-22\n\nFind a useful next step in onboarding, Discovery Briefs, and prototype studies, with clearer limits on what public research and study evidence can show.\n\nWhat changed\n\n- Choose a feedback path when public research finds little\n- Get an input or review action that fits an empty Discovery Brief\n- Keep participant reports and recorded signals distinct from verified study outcomes\n\nMake the next onboarding step clear\n\nWhen public research finds little, Zentrik explains the limit without judging the team's evidence and offers a clear next step.\n\nConnect customer feedback sources during onboarding\n\nOnboarding explains what each source shares and lets the team choose what to bring into Discovery. A source that needs one credential can be connected in the flow; sources that use OAuth return to onboarding after setup, and connecting a source requires settings-admin access.\n\nFind the right input for an empty Brief\n\nGet a useful next step from an empty Discovery Brief, based on whether feedback is new, quiet, pending, or failed. Saved editions and populated reports keep their current design. Viewers do not receive writer controls, and failed checks offer retry.\n\nKeep prototype study evidence grounded\n\nUse more neutral prototype study prompts that separate what participants report from signals the product records and outcomes verified independently. Model output remains probabilistic and human-reviewed; the release's synthetic evaluation does not establish a population result or a customer-study outcome.\n\nMore in this release\n\nOnboarding and research\n\n- Show research phases during consent and explain when a public-feedback run takes longer\n\nDiscovery and planning\n\n- Give initiative records clearer public identifiers and account reconciliation\n- Improve overview navigation, charts, saved filters, and delivery triage\n\nReliability and collaboration\n\n- Distinguish invitation delivery, bounce, suppression, and failure states\n- Limit the searchable Email delivery ledger to super admins\n- Recover supported reads and completed drafts after interrupted work\n\nCanonical page: https://zentrik.ai/releases/v1-30-0\n", "date_published": "2026-09-22T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/onboarding-thin-public-evidence/onboarding-thin-public-evidence-light.png" }, { "id": "https://zentrik.ai/releases/v1-29-0", "url": "https://zentrik.ai/releases/v1-29-0", "title": "v1.29.0: Unified Discovery learning and more reliable connected work", "summary": "Review evidence and insights together, keep Briefs and imports clearer, and recover more safely from API, delivery, and research failures.", "content_text": "Unified Discovery learning and more reliable connected work\n\nv1.29.0 · Released 2026-09-21\n\nReview evidence and insights together, keep Briefs and imports clearer, and recover more safely from API, delivery, and research failures.\n\nWhat changed\n\n- Review evidence and synthesized insights together in Learn\n- Read Briefs with supporting evidence and clearer comparisons\n- Connect Gong with optional OAuth and return to imports\n- Recover assistant choices and connected-work failures with clearer boundaries\n\nBring evidence and insights into one Learn workflow\n\nLearn now brings evidence and synthesized insights together with shared filters, completion actions, and links to related Ideas, Initiatives, and Tasks. Trending uses changes in distinct supporting signals, and evidence-free insights keep their existing completion behavior.\n\n- Review evidence in Discovery: https://zentrik.ai/docs/product/evidence-to-insights\n\nLead Briefs with evidence\n\nWeb and email Briefs now lead with useful findings while preserving source composition, comparisons, and recipient scope. When you review the evidence behind a Brief count, the linked records remain available for inspection.\n\nConnect and recover imports\n\nGong can use optional OAuth when the client ID, secret, and callback are configured; API-key connections remain available when they are not. Jira import receipts and scheduler handoffs remain aligned when completion work overlaps.\n\n- Connect Gong: https://zentrik.ai/docs/integrations/gong\n\nKeep assistant choices and API writes bounded\n\nAssistant choices can appear before a turn finishes, but they are previews until the response is complete and the user confirms; written alternatives remain available. API inputs are validated before persistence, and workspace relation updates commit atomically.\n\nKeep research work moving\n\nOnboarding keeps evidence visible and ends with a defined opportunity, while Study findings handle invalid characters and missing images without hiding the underlying problem.\n\nMore in this release\n\nDiscovery review\n\n- Review evidence and insights in one Learn workflow\n- Open related Ideas, Initiatives, and Tasks from the feed\n- Lead web and email Briefs with supporting findings\n- Preserve source composition, comparison, and recipient scope\n\nConnected work\n\n- Connect Gong through optional OAuth\n- Keep Jira import receipts and scheduler handoffs aligned\n- Recover concurrent GitHub delivery links\n- Keep community evidence scoped to supported participants\n\nReliability and clarity\n\n- Show assistant choices as they arrive and accept written alternatives\n- Validate API values before persistence and commit relation updates atomically\n- Handle missing Study images and invalid finding characters predictably\n- Bound interview-candidate, analytics, and workspace-deletion query work\n\nCanonical page: https://zentrik.ai/releases/v1-29-0\n", "date_published": "2026-09-21T00:00:00Z", "image": "https://zentrik.ai/images/product-demos/discovery-brief-sources/discovery-brief-sources-light.png" } ] }