#Data Engine & Custom Reports

#Grab your files right from a report

What's new: A file attachment column in a report used to say Has attachment: Yes/No, and that was the whole story. Useful, but it meant opening submissions one at a time, or exporting every file and deleting the ones you didn't want, just to track down a single screenshot. Now you can add a column that shows the actual file names as clickable download links.

See it in action:

1. Open a custom report and edit its columns.
2. Pick a file attachment field, and you'll see a new data point named after it: <field name> files.
3. Add it, save, and run the report. Single files show one filename, and multiple files show them comma-separated, each its own link.
4. Click any filename and the file downloads. That's it, no more ZIP archaeology. 🎉

Good to know:

• Your existing Has attachment: Yes/No columns are completely untouched. This is a new, separately selectable data point sitting alongside it, so nothing you've already built changes.
• It's available in the feedback, user, and project report areas.
• You can filter on it too, with a new File name filter: Contains, Does not contain, Is equal to, Is not equal to, Starts with, and Ends with. There's also Is more than and Is less than if you want to filter by how many files are attached.
• Exports come along for the ride. XLSX gives you real clickable Excel hyperlinks, and PDF shows the filenames as plain text.
• Sign-in is always required to download, in every context, including a link clicked from a distributed or emailed report. Nobody gets file access they didn't already have, and existing file security levels still apply.

#Form Engine & Custom Surveys

#Survey presets show up in an order that makes sense

What we fixed: When you added a survey question from a preset, presets borrowed from other surveys in the same project came through in a seemingly random order. They were actually sorted by a behind-the-scenes name you never see on screen. On top of that, every other survey in the project got dumped into one giant undivided section. We've put presets back in proper question order, and each source survey now gets its own labeled group.

See it in action:

1. Open a project with at least two surveys, where one of them has questions saved as presets.
2. In a different survey, add a new element and open the preset drop-down.
3. Presets from other surveys are now grouped under each survey's own name, listed in the same order the questions appear in that survey. 🎉

Good to know: This is how community question bank presets have always behaved, and project surveys just weren't following the same rule. The difference is most obvious on large projects, where a hefty preset library made the old ordering genuinely painful to work through.

#Form Engine & Custom Surveys

#Completed surveys now show the answer you actually picked

What we fixed: On the review screen for a completed survey, a Choices question set to the vertical radio button style could show the first option as the selected answer, no matter which one was picked. The trigger was oddly specific: it only happened when one of the answer choices had the word "checked" somewhere in its text, like "They matched all the times that I checked today." The review screen was working out which option was selected by scanning for that word, so a choice containing it matched itself and won every time. We've rebuilt how the review screen decides, so it reads the answer that was actually stored.

See it in action:

1. Open a survey with a Choices question using the vertical radio button response style, where at least one choice has the word "checked" in its text.
2. Take the survey and select any option other than the first one.
3. Submit the survey, then view the completed survey review screen.
4. The option you selected is the one looking back at you. 🎉

Good to know: This only ever affected how the review screen displayed the answer. The Data Engine, reporting, and the response edit view had the correct answer the whole time, so nothing you've already collected needs a second look. Horizontal radio buttons had the same quirk, and they're fixed too.

#Conditional questions now stay hidden until the matrix question is answered

What we fixed: When a matrix question included N/A as one of its answer choices, any conditional question tied to that row would show up immediately, before anyone had answered a thing. The row looked untouched, but the conditional logic was reading that untouched row as an N/A selection and firing on page load. We've taught it the difference between a row nobody has answered yet and a real, deliberate N/A.

See it in action:

1. Build a matrix question that includes N/A among its answer choices, such as a 1-5 scale plus N/A.
2. Add a conditional question triggered by N/A on a specific matrix row.
3. Preview or take the survey and leave the matrix question unanswered. The conditional stays out of sight where it belongs.
4. Select N/A on that row and the follow-up appears right on cue. Pick a real score instead and it tucks itself away again. 🎉

Good to know: Reopening a submission that was saved with N/A still shows the conditional on load, which is exactly what should happen: that N/A is a genuine answer. Conditionals on standard choice questions, drop-down lists, and checkboxes all behave the way they always have.

Centralized Jira connections are graduating from GoEarly to production sites. If you've ever configured the same Jira instance across ten different projects, this one's for you. A community admin sets up each Jira instance once, and project admins pick it from a drop-down. Your existing project-level Jira setups keep working exactly as they are.

#One Jira Connection, Every Project That Needs It

Jira credentials used to live inside each project's external destination, per feedback type. Same instance URL, same username, same API token, re-entered for every project that needed it. Rotating a token meant touching each project by hand. Now a community admin configures the Jira instance once in the Integration Center, and every project pointing at that connection inherits the change.

Each connection stores an instance name, a base URL, an authentication scheme (basic with a username and API token, or bearer with a personal access token), and the credentials themselves, encrypted at rest. You can set up as many instances as you need, so an engineering Jira and a support Jira can happily coexist.

Every connection card has a Test button. Select it and we call Jira, then show you the list of Jira projects those credentials can actually reach, 30 at a time. It's a quick confirmation rather than a project browser, and it catches the situation that generates the most support tickets today: a token that authenticates perfectly well but lacks the access the sync needs downstream. A successful test shows a green Connected badge. A failure shows a red Connection unsuccessful with the reason in plain language, like Token expired or Could not reach Jira at this URL. Save credentials without testing them and the card sits at Untested until you do.

Each card also shows a Used by # Centercode projects count that expands into the actual list, so you can see who depends on a connection before you change it. A footnote records who edited the connection last and when. If you remove a connection that's still in use, we tell you how many Centercode projects will break and let you decide whether to continue. Those projects reset to Select and their Jira sync fails on the next attempt, which is the signal for their admins to pick something new.

In a project, the authentication fields on the Jira external destination are replaced by a Connection source drop-down. It lists your community's connections alphabetically, with Project-specific connection at the bottom. Choose a community connection and the instance details display read-only, so you go straight to the JSON mapping without hunting down a token. For a one-off integration, like a contractor's Jira account, choose Project-specific connection and enter the base URL, authentication scheme, and credentials the way you do today.

Switching between the two is safe in both directions. When you move a project from its own credentials to a community connection, the original credentials stay put behind the scenes, obfuscated. If that community connection later disappears, flip back to Project-specific connection and everything you entered is still there waiting.

Important note: Nothing migrates on release day. Every existing Jira external destination lands on Project-specific connection with its current credentials intact, and Project-specific connection isn't going anywhere. Opt in to a community connection on your own schedule, or don't. Your JSON mapping and field configuration behave identically either way. For the full setup walkthrough, see our Guide to Centralized Jira Connetions.

---

#Dynamic Tags Get the Same Picker as Everywhere Else

The old inline drop-down for inserting dynamic tags into the JSON editor is retired. The dynamic tag button now opens the same search modal you already use elsewhere in the platform. Park your cursor in the JSON, open the modal, add the tags you need, close it, then move them where they belong.

Dynamic tags are a one-and-done setup task, so the extra click buys you a picker that works the same way everywhere.

#Onboarding/Recruiting

#Joining a project with SSO now works from long recruitment links

What we fixed: If you had a URL recruitment set up and someone signed in with SSO right from that recruitment page, they'd get bounced to our "no page exists at your destination address" error page instead of being joined to the project. Not exactly the warm welcome we were going for. Behind the scenes, the link they were sent to after signing in wasn't being recognized properly because of its length, so the whole join stalled out before the application survey ever loaded. We've corrected the routing, so those links resolve the way they should have all along.

See it in action:

1. Head to a project with a URL recruitment set up with a key over 40 characters, the kind with a custom link like yoursite.com/your-recruitment-name.
2. Open that recruitment link in a fresh browser, or an incognito window, so you're starting signed out.
3. Sign in using SSO from the recruitment page.
4. You'll land right where you should, moving into the project's application survey instead of an error page. 🎉

Good to know: This showed up most often on recruitments with longer, multi-word keys, but the fix covers all of them. If anyone reported a "bomb error" when trying to join through a recruitment link, this was almost certainly the culprit, and it's worth letting them know they can try again.

#Your "send without an opportunity link" confirmation now sticks

What we fixed: When you send a recruitment email that uses the %OPPORTUNITYURL% tag, you have to check Allow email to send without an opportunity link to move forward. Trouble was, if you then hit the recipient-count validation error on the preview page, that checkbox quietly forgot you'd ever confirmed it. You'd fix the recipient count, hit send again, and get blocked by an opportunity-link error, with no way forward except backing up a page and re-checking the box. We've fixed that, so your confirmation carries through.

See it in action:

1. Create or open a recruitment email that uses the %OPPORTUNITYURL% tag with a custom footer.
2. Check Allow email to send without an opportunity link and continue to the preview page.
3. Click send without confirming your recipient count first, so you trigger the validation error on purpose. (Yes, we're asking you to break it. Live a little.)
4. Now confirm your recipient count and send again. It goes through, with no second detour back to re-check that box. 🎉

Good to know: Nothing about how the email itself sends has changed here. This is purely about the confirmation sticking around so you're not stuck in a loop on the preview page.

#Integrations

#App store release automation now runs again

What we fixed: If you had TestFlight or Google Play automation set up, it could quietly refuse to do its job: either stalling out behind the scenes, or never kicking off at all while the settings screen happily showed it as enabled. We've fixed the setting it was misreading, so automation now runs when it says it's on.

See it in action:

1. Head to Community > Integrations and open your TestFlight or Google Play integration.
2. Check the linked projects list to confirm your project is connected and automation is switched on.
3. Push a new build up to the store, then pop over to the integration's log.
4. You should see the build get picked up and processed, with no mystery silence. 🎉

Good to know: The sneakier version of this one left no trace at all. Automation simply never started, with nothing in the log to explain why. If you'd given up on an automation setup because it seemed to do nothing, it's worth another look.

#Webhook macros that pull in outside data no longer crash

What we fixed: Running a data-pull webhook macro (a Salesforce lookup, for example) could fall over if the other system came back with nothing at all. From a user's User Summary page you'd get a script error, and from User Management the progress bar would just spin forever without ever telling you why. We now handle those responses gracefully.

See it in action:

1. Open any user's User Summary page in a community that has a webhook macro configured.
2. Click the macro from the left-hand macro links and run it.
3. It completes cleanly, with no error popup this time.
4. Try the same macro from User Management. The progress bar actually finishes instead of spinning into eternity. 🎉

Good to know: If the external system genuinely does return nothing, the macro now fails cleanly with a visible error rather than taking the whole run down with it, so there's something concrete to look at instead of a silent hang.

#Test Planning

#Product Verification Stays Published Like It Should

What we fixed: Product Verification was quietly flipping to an unpublished state after an admin opened one of its associated macros and came back. Since Product Verification acts as a hard gate, that meant everyone in the trial got stuck until someone noticed and republished it. We fixed it so Product Verification holds its published state.

See it in action:

1. Open a project with Product Verification published and gating trial access.
2. Click into one of its associated macros, then head back to the Product Verification page.
3. Product Verification is still published, right where you left it. 🎉

Good to know: Product Verification now only unpublishes when you choose to do so yourself.

#Email

#Large Email Resends No Longer Hit an Error

What we fixed: Resending a big batch of community-level email at once (think 100 at a time) could throw a server error. We fixed it so large batches go through without a hitch.

See it in action:

1. Head to a community-level email with a hefty queue.
2. Resend a large batch, for example 100 recipients.
3. It sends without a hitch, no need to break it into smaller chunks. 🎉

#Form Engine & Custom Surveys

#AI-Generated Surveys Point Conditionals at the Right Question

What we fixed: When AI built or imported a survey with conditional logic, the branching rules landed on the wrong question, consistently one spot off. That meant admins had to fix each one by hand. We tightened up the logic so conditionals now point to the question they were actually meant for.

See it in action:

1. Use AI survey creation to generate a survey that includes conditional (display or branching) logic.
2. Review each conditional against your element list.
3. Everything points where it should, no manual cleanup needed. 🎉

Good to know: We also tightened up how character limits are handled during AI survey creation as part of this fix.

#Feedback

#Summary Prompt Enhancements No Longer Disable AI Reply Generation

What we fixed: Adding a prompt enhancement to summary generation had a side effect: it switched off AI reply generation and wiped out its custom prompt. We fixed it so these two settings stay independent of each other.

See it in action:

1. Turn on both AI reply generation and summary generation in Feedback collaboration.
2. Add a prompt enhancement to summary generation.
3. Your AI reply generation setting and its custom prompt stay exactly as you left them. 🎉

#AI-Enhanced Feedback Keeps Its Formatting

What we fixed: When AI enhancement ran on a multi-line or rich-text feedback element, the enhanced version flattened everything into one run-on paragraph, losing line breaks and italics. We fixed it so the enhanced view keeps the structure the original author intended.

See it in action:

1. Open a feedback item with AI enhancement enabled on a rich-text element that has line breaks and italics.
2. Toggle to the AI-enhanced version.
3. Line breaks stay as separate lines and italics stay italic. 🎉

#Onboarding/Recruiting

#Revoke Pending Invitations Through the Onboarding API

What we fixed: The onboarding (user import) API only worked one way. You could push users into a pool, but there was no way to pull an unused invitation back out without going into the platform directly. We added a revoke counterpart to the add endpoint, so integrations can cancel a pending invitation programmatically.

See it in action:

1. Add a user to an onboarding pool via the API, as you do today.
2. Call the new revoke endpoint with the same email you used to add them.
3. That pending invitation can no longer be consumed. The person only gets back in through a fresh invitation. 🎉

Good to know: This only affects pending invitations, anyone who already joined is unaffected. The revoke endpoint requires the encoded email as a parameter to succeed.

#Data Engine & Custom Reports

#SSO Source and Integration ID Line Up in Reports

What we fixed: In custom reports, the SSO source and SSO integration ID columns sorted their multi-value entries in different orders, making it tricky to tell which ID matched which source. We fixed it so both columns order their entries the same way.

See it in action:

1. Create a custom report showing SSO source and SSO integration ID.
2. Check a member with more than one SSO source.
3. The sources and their IDs line up in the same order across both columns. 🎉

#User Accounts

#SSO sign-in now works even when your login provider skips the email

What we fixed: If you invited someone to a project and they signed in through SSO, they could get stuck at the door when their identity provider didn't hand over an email address. Even when their real email matched the one you invited, Centercode had nothing to match it against, so they were left standing on the welcome mat. We taught Centercode to reconcile the invite's email with the authenticated SSO identity, so these folks can sail right through.

See it in action:

1. Invite a user to a project by their email address.
2. Have them sign in through your SSO provider, even one that doesn't return an email in its response.
3. They complete sign-in and land in the project. No more dead ends! 🎉

Good to know: This was a rare edge case, but a total roadblock when it hit. Affected invitees had no way in, and there was no workaround on our side. If your identity provider keeps email addresses close to the vest, you're covered now.

#User Agreements

#Signed agreements now get their PDF copies

What we fixed: If your community uses a custom primary domain (your own branded URL), signed user agreements weren't generating downloadable PDF copies. This meant the PDF and Download all agreements buttons went missing entirely. We tracked the cause to how the archiving job builds its capture URL, and agreements now get properly archived no matter what domain you're on.

See it in action:

1. Go to Community > Agreement center and search for a user who has signed an agreement.
2. Hover over their agreement row. The PDF button is back where it belongs.
3. Open the agreement and you'll find Download as PDF at the bottom, plus Download all agreements ready to go. 🎉

Good to know: Agreements that failed to snapshot before this fix will need a backfill on our end to catch up. Reach out if you're missing older PDFs and we'll get them regenerated.

#Integrations

#Jira pending status now checks the right team

What we fixed: Feedback tickets sometimes showed a "Jira pending" status when they shouldn't have. If one team had a Jira external destination set up on a status, another team using that same status got tagged as pending too, even without any Jira integration of its own. We updated the check to pay attention to which team is actually involved.

See it in action:

1. Set up a Jira external destination on one team's status (say, your PM team's New).
2. Have a different team's matching status (like a participant New) route normally, with no external destination.
3. Submit a new ticket and view it. The status stays clean, with no stray "Jira pending" tag where it doesn't belong. 🎉

#Slack stops double-posting your feedback

What we fixed: When AI feedback features were switched on, Slack notifications fired again every time a feedback item was edited or had its status changed, not just on the original submission. We applied the same one-and-done guard the standard path already used, so Slack posts exactly once.

See it in action:

1. On a project with Slack integration and AI feedback summaries active, submit a new feedback ticket.
2. Watch your Slack channel light up with the one post you'd expect. 🎉
3. Reopen that ticket, change its status, and save. Your Slack channel stays quiet.

#Feedback

#Impact score widget text stays visible on light themes

What we fixed: On the feedback view, the impact score widget's project points, community points, and user name borrowed the Text links and header buttons theme color. If that color happened to be white or close to it, the text turned white-on-white and disappeared. Those values now use the same readable color as links inside a feedback description, no matter your theme.

See it in action:

1. Set your theme's Text links and header buttons color to white or something very light.
2. Open a feedback item and look at the impact score widget.
3. Project points, community points, and the user's name are all clearly visible. 🎉

Summer's here, and so is a release we've been looking forward to. The headliner is Required on Duplicate, graduating to live servers after its run on GoEarly. Duplicates used to be a mechanism for exclusively adding weight to existing feedback. Now they can carry real context forward into your views, reporting, and exports. Over on GoEarly, we're previewing Centralized Jira Integration, which lets you set up your Jira connections once at the community level and share them across projects. Before we get into the features, though, here's a change behind the scenes that we're genuinely excited about.

#An exciting change in how we ship

We've changed how we build and deliver Centercode behind the scenes, and it's a change we're excited about, because you're the one who benefits. What it means for you: we can get new capabilities into your hands sooner, react faster when you need something adjusted, and roll out improvements without interrupting your work. It's the kind of thing you'll feel more and more as we go.

#✨ New Features ✨

#Required on Duplicate

When a participant runs into something another participant already reported, they have two quick ways to say "me too": voting on an existing issue, or matching a predictive suggestion as they submit. Both paths used to do the same thing behind the scenes. They added one to a counter. You knew more people hit the issue, but not what was different about each encounter: their device, their build, the time it happened, the steps that led there.

Required on Duplicate closes that gap. You can now flag specific form elements as required on duplicate, so that when a participant claims an existing issue, they're prompted to fill in just those fields before the duplicate is recorded. A focused popup appears with only the fields you've marked, nothing else, so it stays quick for the participant while capturing what your team actually needs to triage.

That captured context travels with the duplicate. Each "me too" now creates a real duplicate submission, linked to the original, with the participant's data attached. From there it flows into your comparison view, reporting, and exports, the same as any other feedback. You're triaging from what people actually reported, not from a count and a guess.

Setting it up is per element and entirely up to you. Inside a feedback type, open the form element you want to collect on a duplicate and turn on required on duplicate. Good candidates are short, high-signal fields like time, device, build, and steps to reproduce. Eligible field types include text inputs, date and time pickers, single and multi-select drop-downs, and rating elements. File uploads, AI-enhanced fields, and read-only elements stay out of the duplicate flow to keep the popup fast. If a participant cancels the popup, no duplicate is recorded, so only complete submissions count.

A few things worth knowing. Duplicates collected before you turned the setting on will show blanks for those fields, since the data wasn't requested at the time. The same participant can log more than one occurrence on the same issue, which is intentional. And if you ever remove a duplicate relationship, the context you collected stays put. Read more here!

#💡 GoEarly Features 💡

#Centralized Jira Integration

If you run Jira integrations across more than a couple of projects, you know the drill. Every project, every feedback type, the same credentials maintained by hand. When an access token expires, you go update them one at a time, and there's no single place to see which projects share a connection or whether a key still has access.

Centralized Jira Integration, now previewing on GoEarly, lets you manage your Jira connections once at the community level instead of project by project. Set up a connection a single time, with its name, base URL, authentication type, and credentials, and share it everywhere it's needed. Each connection shows which projects rely on it and when it was last edited, so you finally have a clear picture of what's connected to what.

Project-level destinations gain a connection-source choice. Point a destination at a shared community connection, or keep using project-specific credentials exactly as you do today. Nothing about your existing setups changes unless you decide to move them, and your credentials are preserved through every step: switching to a shared connection, having one removed, and recovering.

Our favorite part is the built-in Test action. It checks a connection and shows you which Jira projects it can actually reach, which catches the most common snag we see in support: a key that connects just fine but quietly lacks access to the project you're aiming at. Instead of guessing, you get a straight answer.

Want a first look? Sign up for GoEarly now!

#Feedback Workflows

#Notify new owner emails now reach the whole team

What we fixed: When feedback ownership moved to a team, the Notify new owner email could quietly skip some members. If one person on the team had opted out of unessential email, everyone listed after them was dropped and never got the notification. We rebuilt the recipient logic so that only the person who opted out is skipped. Everyone else on the team gets the email as expected.

See it in action:

1. Set up a feedback workflow that notifies a new owner, with a team as the new owner.
2. Have one member of that team opt out of unessential email.
3. Trigger the ownership change and confirm every eligible team member receives the notification. 🎉

#Integrations

#Mapping users to Jira fields now works reliably

What we fixed: Sending feedback to Jira could fail when a user field like Reporter or Assignee was mapped. The Insert JSON data helper was building those fields in a format modern Jira Cloud no longer accepts, so Jira rejected the affected tickets. We updated the helper to use the account-based format Jira now requires, and the fix covers Reporter, Assignee, and any user-type custom field.

Good to know: User fields in Jira are identified by account ID, so you'll need to grab the account ID from Jira and drop it in where the tag or placeholder would normally go. Our Jira integration FAQ walks through where to find it.

#Data Engine & Custom Reports

#Custom reports can now include duplicate feedback

What we fixed: Custom reports built on a feedback data source left duplicate submissions out, so there was no clean way to roll them into your reporting. We added an Include duplicates option to feedback report elements, so you can choose to fold duplicate feedback into a report alongside everything else.

See it in action:

1. Open a custom report and add or edit a feedback element built on a feedback data source.
2. Enable Include duplicates.
3. Save the report and confirm duplicate feedback is now included. 🎉

Good to know: The option lives on the individual report element, so you can include duplicates in one element while leaving the others as they are.

#Feedback

#Workflow changes now save reliably on large team projects

What we fixed: On projects with a lot of teams, saving a change to a feedback workflow could fail with an error and refuse to go through. The workflow page was sending far more data than it needed on every save, and past a certain team count that payload got large enough to be rejected before the save could even run. We rebuilt the save so it only sends what it actually needs, which means changes go through cleanly now even on the most team-heavy projects.

See it in action:

1. Open a project with a large team structure and go to its feedback workflow settings.
2. Edit a workflow step. For example, set a state to allow New with no team change.
3. Save the change and it goes through without an error. 🎉

Good to know: This also makes the workflow configuration screen a little snappier, since there's less data moving back and forth behind the scenes.

#Feedback recommendations stay hidden when a type isn't accessible

What we fixed: When a feedback type was turned off for everyone, the recommendation lightbox and its link could still appear, and selecting that link sent participants to an access-denied page. We updated the prompt so it checks whether the participant can actually submit to the linked type first. If they can't, the prompt stays hidden and no one lands on a dead end.

See it in action:

1. Set a feedback type's access to off for all users.
2. As a participant, go through the flow where the recommendation lightbox would normally appear.
3. Confirm the lightbox and its link no longer show, so there's no access error. 🎉

#Integrations

#Removing a macro or listener now clears the External Data fields it created

What we fixed: When you removed a Webhook macro or an External destination listener, the External Data fields it had created were left behind. Those orphaned fields still showed up as view columns and dynamic tags, with no way to clean them up on your own. Removing the macro or listener now also removes the External Data fields it created, so they no longer linger in your views or tag pickers.

See it in action:

1. Create a macro with a webhook and data return that produces an External Data field, then run it once so the field appears.
2. Remove the macro.
3. Open a view or the dynamic tag picker and confirm the leftover field is gone. 🎉

Good to know: This kind of cleanup used to require a request to our team. It happens on its own now.

#Failed Jira pushes now show a clear error instead of staying pending

What we fixed: When a push of feedback to Jira failed, the tag stayed on Pending indefinitely, so there was no easy way to tell the push had actually failed. Failed pushes now move into a clear error state that links straight to the integration log, so admins can open it and diagnose what went wrong. Pending still shows while a push is in progress and updates on its own once the Jira issue is created.

See it in action:

1. As an admin, open a feedback ticket whose Jira push has failed.
2. Notice the tag now shows a clear error state instead of Pending.
3. Select the error to open the integration log and diagnose the issue. 🎉

Good to know: This error state is admin-only. If you display Jira tags to participants, they'll only ever see a resolved link, never a pending or errored one.

#Data Engine & Custom Reports

#External Data filters now include records with no value

What we fixed: The Is not equal to and Does not contain filters on External Data fields were skipping records that had no value for that field. Because an empty field has nothing to compare against, those records were dropped from the results, which in some cases made the filters return nothing at all. We corrected the comparison so an empty value now counts as not equal to and does not contain, and those records show up where you'd expect.

See it in action:

1. Open a User or Feedback view in a project with an External Data field that's empty on most records.
2. Add a filter on that field using Is not equal to or Does not contain.
3. Confirm records with no value for the field now appear in the results. 🎉

#New Features

#Embed Interactive HTML Prototypes Right in Your Content

What we fixed: Want to drop a self-contained HTML prototype into your content? You can now upload it as an External Embed asset and reference it inline, and it'll render in its own isolated frame. No more crashes on save or rendering hiccups from pasting big chunks of prototype HTML into the content body.

See it in action:

1. Head to the External Embed header-assets block and upload your HTML file.
2. Reference it inline in your embed code.
3. Publish and watch your prototype come to life in a tidy, sandboxed frame. 🎉

Good to know: This works best with a single self-contained file. A split HTML + CSS + JS set won't resolve its relative links, so bundle it all into one file for the smoothest results.

#Feedback

#Your Display Name Stays Your Display Name in AI Summaries

What we fixed: When the AI summarized a feedback discussion, it was sometimes calling participants by their real first name (pulled from their profile) instead of the display name shown right there on the ticket. A little confusing, and not great for folks who set a display name specifically to control how they show up. We tracked it down to the summary reading names straight from the user profile and rebuilt it to respect the ticket. AI summaries now stick to the display name, just like the rest of the platform.

See it in action:

1. Open a feedback ticket where the original poster has a display name that's different from their profile first name.
2. Generate the AI summary on the discussion.
3. Check the name in the summary. It'll match the display name on the ticket, no surprises.

Good to know: This keeps name-masking consistent everywhere, so participants always show up the way they expect to.

#Data Engine & Custom Reports

#Excel Exports No Longer Trip Over Long Element Names

What we fixed: Excel caps sheet names at 31 characters, so we shorten long report element names to fit on export. If that shortened name happened to land on a character Excel doesn't allow at the end (an apostrophe, for example), the whole export would fall over. We've smoothed that out, so long names won't break your export anymore.

See it in action:

1. Build a report with an element that has a long name (bonus points for one with an apostrophe right around the 31st character).
2. Export to Excel.
3. Watch it export cleanly instead of throwing an error. 🎉

#Integrations

#Email Jira Links Now Match Your In-Site Links

What we fixed: If you set up a custom Public Link Base URL for an external Jira destination, the in-site "pill" links used it, but the "Jira link" dynamic tag in workflow and feedback macro emails was still pointing at the original Jira URL. The same record could send people to two different places depending on whether they clicked in-site or in an email. Now both surfaces use the same configured URL.

See it in action:

1. In an external Jira destination, set a Public Link Base URL that's different from the technical base URL.
2. Check the Jira "pill" link on a feedback record in-site.
3. Trigger a workflow or macro email with the "Jira link" dynamic tag for the same record. The email link now matches the in-site one.

#Integrations

#Jira Tags Now Show Up Right Away

What we fixed: Previously, when you moved a piece of feedback into a status linked to a Jira destination, the Jira tag could take a couple of minutes to appear, long enough that folks figured something was broken and started toggling statuses to force it. Now the tag shows up immediately with a friendly Pending… label and updates itself the moment the Jira issue is created. No refreshing, no second-guessing!

See it in action:

1. Open a feedback item in a project that has a Jira destination configured.
2. Change its status to the one tied to your Jira destination and save.
3. Watch the blue Jira tag appear right away with a Pending… note. It'll swap itself out for the real Jira ID once it's ready. 🎉

Good to know: This is especially handy on projects using AI summaries or discussions, where the Jira ID generation can take a little longer behind the scenes.

#Feedback

#AI-Assisted Feedback Now Submits with Just a Title or Description

What we fixed: AI-assisted feedback could be blocked from submitting when the form contained only a title and/or description with nothing else marked as a field. The front end allowed it through, but the back end was holding it back. We've adjusted the check so a minimal title-and-description form submits cleanly.

See it in action:

1. Set up a feedback type with no required form elements beyond title and description.
2. Turn on AI-assisted feedback submission.
3. Submit a ticket without modifying the form. The submission goes through cleanly. 🎉

#User Accounts

#A Little Extra Security When Removing Your Account

What we fixed: To help keep accounts safe, local-login users are now asked to confirm their password before removing their own account. It's a quick extra step that makes sure it's really you before anything disappears.

See it in action:

1. Head to your account removal (opt-out) page while signed in with a local login.
2. You'll notice a new password field on the confirmation step.
3. Pop in your password to confirm and you're done. 🎉

Good to know: This only applies to the logged-in self-removal page for local accounts. SSO sign-ins are unaffected and won't see a password prompt.

#Closed Phases, Reopened: Test Planning Gets More Forgiving

Test planning picks up three new moves this month, headlined by the big one: you can now reopen a completed phase and pick up data collection where it left off, without losing a thing. Date edits on past phases and single-day phase creation round out the trio. And a new GoEarly preview lets admins capture real per-participant context every time someone votes "me too" or matches a duplicate. Want a sneak peek? Sign up for GoEarly now!

#✨ New Features ✨

#Three Test-Planning Moves That Ease Long-Standing Friction

A completed phase can now be reopened with one action. The new reopen control on closed phases flips the phase's status back to active and reactivates its surveys and activities, so participants can submit data again. Every submission, response, and attachment from the prior window stays intact. The reopened phase resumes its identity rather than starting over. If you've ever spun up new artifacts to run a follow-up round (and lost the historical thread doing it), this one's for you.

Date edits on past phases let you fix what was wrong without rewriting history. Deactivated phases now show the same date-edit controls as active and pending ones. Correcting a typo, retroactively labeling a phase to match what really happened, or aligning a phase to an external timeline is now a self-service edit. Changing a deactivated phase's dates doesn't reactivate it. The phase stays deactivated unless you also choose to reopen it.

Single-day phases finally just work. Set the start date to match the end date, and the platform accepts it cleanly. The timeline view renders the phase visibly rather than as a zero-width sliver, so single-day events stay recognizable next to multi-day ones. Useful for one-day kickoffs, focused field tests, or pinpoint survey distributions that didn't quite belong in a longer window.

All three new affordances gate on the same permission as creating a phase, so there's no new role surface to teach. If you can create a phase, you can use any of these. The phase status model stays the same (active, deactivated, pending). End dates still don't auto-close phases, and phases still can't overlap.

---

#💡 GoEarly Features 💡

#Capture Real Context on Every Duplicate Vote

Today, when a participant votes "me too" on existing feedback or matches a predictive duplicate mid-submission, the platform records the occurrence but nothing else. This GoEarly preview lets you change that. Admins can require a short set of fields (think time stamps, device IDs, or build versions) every time someone claims an existing issue, captured per participant before the occurrence is recorded.

Most useful for crash, stability, and intermittent-issue investigations, where the "when, on what, after what" is exactly the data that turns a vote count into a triage signal. Full setup details and the in-platform walkthrough come with the production release. Want to try it in a project now?

Sign up for GoEarly now!

#Data Engine & Custom Reports

#Custom Report Exports Render Cleanly in Desktop Excel and Numbers

What we fixed: An index page in a custom report export would render with an unintended black background fill when the workbook was opened in desktop Microsoft Excel or Mac Numbers. The same file rendered cleanly in Google Sheets and Excel Online because those renderers normalize the styling differently. We tracked it down to the cell styling emitted on the index page during export and tightened it up, so the workbook now renders consistently across all spreadsheet apps.

See it in action:

1. Export a custom report that includes an index page.
2. Open the resulting workbook in desktop Microsoft Excel or in Mac Numbers.
3. Locate the index page sheet and confirm the cells render without an unintended black fill. 🎉

#User Accounts

#SAML SSO Logins Continue Working After Attribute Changes

What we fixed: SAML SSO logins could fail after an upstream attribute (such as alias or email) was changed on a user account. We tracked the issue to how the platform reconciled the updated identity, and we've rebuilt that step so logins keep working through identity changes.

See it in action:

1. Sign in to a community configured for SAML SSO.
2. Update the user's alias or email at the identity provider.
3. Sign back in. Authentication completes successfully. 🎉

#Data Engine & Custom Reports

#Report Elements Built on Archived Data Now Reopen Correctly for Editing

What we fixed: If you built a report element that referenced archived view data and saved it, reopening the element for editing showed an empty selection panel. The data was saved correctly on the server, just not rendered when you came back to edit. We aligned the view hydration with how filters already work, so your saved selections reappear no matter how the archive toggle is set.

See it in action:

1. In a community with archived form fields, open Project → Reports and create a new report.
2. Add a report element. In the Choose your data panel, click View archived data.
3. Select a few archived data points and save the element.
4. Click back into the saved element. Your archived selections appear in the right-hand list. 🎉

Good to know: The left-side source tree still defaults to active data. Flip View archived data to see, add, or remove archived items in the tree.

#Feedback

#Comments Section Now Loads Reliably on Feedback Views

What we fixed: The comments section on a feedback view could intermittently fail to appear, throw an error, or show the wrong page content after a refresh. It was an elusive one, but we tracked it down and got it sorted.

See it in action:

1. Navigate to a feedback view.
2. Refresh the page.
3. The feedback comments section now loads correctly — no errors or wrong page content.

#Project Homepage

#Homepage Load Times Improved for Large Projects

What we fixed: If your project had a lot of users, feedback types, and submissions, the homepage could get sluggish or even time out before it finished loading. The query behind feedback button counts was doing far more work than it needed to. We rebuilt it to fetch only what the page actually uses.

See it in action:

1. Navigate to the project homepage for a project with large numbers of users, feedback types, and submissions.
2. The page now loads quickly instead of lagging or timing out. 🎉

#Form Engine & Custom Surveys

#Screen Readers Now Correctly Announce Rating-Scale Choice Labels

What we fixed: VoiceOver and other screen readers were having a rough time with rating-scale choices in surveys. Labels were showing up garbled, unlabeled, or nested inside confusing groups, which made forms genuinely difficult to navigate with assistive technology. We dug into how each choice was being announced and corrected the labeling and grouping across the board.

See it in action:

1. Navigate to a survey containing rating-scale choices.
2. Use VoiceOver or another assistive technology to navigate those choices.
3. Labels are now read correctly instead of being announced as unlabeled or grouped incorrectly. 🎉

#User Accounts

#SSO Account Creation No Longer Blocked by Display Name Setting

What we fixed: When Disable display name change was enabled in Community > User account settings, users creating an account for the first time through SSO could get stuck mid-setup and couldn't complete account creation. We've fixed the SSO path so new users can set a display name and get through setup without hitting a wall.

See it in action:

1. Enable Disable display name change in Community > User account settings.
2. Have a user create an account via SSO.
3. The user can now set a valid display name to complete account creation.

#Cleared Usernames Can Now Be Updated When Display Name Lock Is On

What we fixed: The Disable display name change setting is meant to stop users from swapping out a username they already have — not to block users who don't have one at all. But if an admin had cleared a user's username, that user would hit a confusing validation error when trying to set a new one. We've tightened up the logic so the lock only applies when a username already exists.

See it in action:

1. Enable Disable display name change in Community > User account settings.
2. Clear a user's username as an admin.
3. Have the user attempt to set a new username.
4. The validation now allows the new username to be saved. 🎉

April’s release is a big one. Two features that have been cooking in GoEarly are officially out of the oven and better than ever, shaped by real feedback from our GoEarly community along the way. AI Feedback Submission and AI Element Enhancement are now live for everyone, and together they represent the biggest shift in how feedback gets collected on the Centercode Platform in a long time. Read on for the full picture, and check out what’s headed into GoEarly next!

#✨ New Features ✨

#The Feedback Form, Reimagined

Structured feedback has always required structure from the people submitting it. Pick a type, fill out the fields, repeat. It works, but it puts the burden on participants to understand your project’s feedback architecture before they’ve even written a word. AI Feedback Submission changes that entirely.

With this feature turned on, participants get one button and a blank text box. They describe what they experienced in plain language, and Ted takes it from there. He figures out the feedback type, fills in the form fields, and checks for duplicates so participants can vote on existing items instead of creating noise. Before anything is finalized, they get a five-minute review window to check Ted’s work and make any edits they want. The form is still there. The structure is still there. Participants just don’t have to think about either of them anymore.

For admins, AI Feedback Submission is a project-level toggle. Turn it on, and every submission in that project flows through the new experience. A Show Original option in the feedback list lets you pull up exactly what the participant wrote before Ted got involved, so nothing gets lost in translation.

This one has the potential to change how your whole community thinks about submitting feedback. Get started in your community settings, and visit the knowledge base for the full setup guide.

Becomes:

---

#Forms That Fill Themselves In

Setting up a great feedback form takes real effort. AI Element Enhancement means that effort goes even further.

With this feature, you can mark specific form fields as AI-generated. After a participant submits, Ted reads the submission and automatically populates those fields. Participants fill out the same form they always have. There's no new flow to learn, no extra steps, and no AI indicators on the form itself. The enhancement happens quietly in the background after they hit submit.

For each AI-enabled field, you choose how Ted handles it. Some fields you'll want Ted to improve on what the participant wrote. A rough, vague title becomes something clean and specific. Other fields you may not want participants filling in at all. Ted can generate those entirely from context, pulling from whichever other fields you point it at. Either way, you write a custom prompt per element that tells Ted exactly what you're looking for, so the output stays consistent across every submission.

When writing your prompt, you have two tools for feeding Ted the right context. Dynamic tags let you pull in data from other fields on the form. For example, a prompt that generates a title can use a dynamic tag to reference the participant's description field, so Ted always has something concrete to work from. The Include this answer in AI prompt checkbox works differently: it tells Ted to include the current field's own value in the prompt. Use this when you want Ted to clean up or improve what the participant wrote in that field directly, no dynamic tag needed. The two work well together, but they do different things.

On the admin side, fields that Ted enhanced still show the participant's original answer alongside the AI-generated version, so you always have both for reference. Find AI Element Enhancement in your community settings and configure it field by field in your form editor. The knowledge base article walks through the full setup.

---

#💡 GoEarly Features 💡

#Phase Scheduling is Getting a Whole Lot More Flexible

If you’ve ever missed a phase end date by a few hours and spent more time than you’d like moving features around, this one’s for you. We’re giving test planning phase scheduling a serious upgrade in GoEarly!

Admins can now create single-day phases, edit dates on phases that have already passed or been deactivated, and re-open completed phases to resume data collection without losing anything that was already captured. No more scrambling. No more moving features around because the calendar didn’t cooperate.

Want early access? Sign up for GoEarly now!

#Show Jira ID to Testers!

What we added: Testers had no way to see the Jira Issue ID associated with their feedback, which made it tough to communicate about unresolved bugs using a shared reference point. Now there's a toggle for that! **See it in action:

1. Go to Feedback properties of any feedback type

2. Enable the new "Show Jira ID to testers" toggle (it's off by default)

3. Testers will now see the Jira ID as plain text in the right-hand menu of the feedback view page **Good to know: The ID is shown as text only (not a clickable link). PMs and Support users (with the Manage feedback and replies role) continue to see the existing Jira link in the title area regardless of this setting. **Good to know: The ID is shown as text only (not a clickable link). PMs and Support users (with the Manage feedback and replies role) continue to see the existing Jira link in the title area regardless of this setting.

#Community and Project Homes

"Restore Content" Button Now Persists After Page Refresh**What we fixed: When testers dismissed content on a project or community homepage and then refreshed the page, the "Restore content" button would disappear. The only way to get content back was to ask an admin to create new content — not ideal! The restore button now sticks around as long as there's dismissed content to bring back. **See it in action:

1. As a tester, visit a project or community homepage that has dismissible content

2. Click the X to dismiss a content block

3. Confirm the "Restore content" button appears

4. Refresh the page — the button should still be there, ready to bring your content back

#Feedback Management

Jira Ticket IDs No Longer Visible to Participants**What we fixed: Participants were seeing Jira ticket IDs and links when viewing feedback — that's info meant for team members only! This snuck in with a recent UI update. Those Jira references are now hidden from participants again, just like they should be. **See it in action:

1. Navigate to a feedback type that has a Jira integration configured

2. Assume a participant who has submitted feedback that was sent to Jira

3. Open the feedback item — you should see no Jira ID or links anywhere in the ticket view (unless the new setting above is enabled for that project)

#Integrations

Jira Integration Now Supports Separate Public Domains**What we fixed: Some customers use a different public-facing Jira URL than their API domain (e.g., their team browses Jira at one address, but the API lives at another). Our integration previously assumed these were the same. We've added support for configuring a separate public domain so Jira links generated by the platform point to the right place. **See it in action:

1. Navigate to a Jira external destination configuration

2. Look for the new public domain field

3. If the customer's public Jira URL differs from their API URL, enter it here — links in the platform will now use this domain

#Data Engine & Custom Reports

#Report Creation From Feedback Management Page Now Works Properly

What we fixed: When clicking "Create a report" directly from the Feedback management page, the second series and filter dropdowns weren't populating correctly.

See it in action:

1. Navigate to the Feedback management page
2. Mouse over a feedback type and click "Create a report"
3. The second series and filter dropdowns now populate correctly

#New CDE Columns: Phase Started and Phase Ended

What we fixed: The CDE Feature Engagement section was missing phase start and end date columns, so there was no way to see when a phase began or ended in data exports. Two new columns, "Phase started" and "Phase ended", are now available.

See it in action:

1. Navigate to a data export that includes CDE Feature Engagement data
2. Look for the new "Phase started" and "Phase ended" columns
3. All features within the same phase will show the same dates — features with no phase association will show blank values

#Integrations

#File Duplication Fixed for API-Uploaded Feedback Attachments

What we fixed: If a feedback ticket was created via an external source with multiple files uploaded through the API, those files could appear duplicated whenever the ticket's status was changed by an external listener macro. Even worse, the duplication compounded with each subsequent status change, so files could multiply quickly.

See it in action:

1. Create a feedback ticket via an external source (e.g., Jira sync)
2. Upload multiple files to the ticket through the API
3. Trigger a status change via an external listener macro
4. Files now display correctly without any duplication, no matter how many status changes happen 🎉

#Onboarding/Recruiting

#Recruiting Landing Page No Longer Creates Broken Invite Types

What we fixed: Editing the "Invited manually via teams" invite type to add a landing page was creating a recruitment in a weird, half-broken state. The recruitment would show only the landing page, and any invites added to it would become invisible (though they were still technically functional and eating up invite URLs behind the scenes).

See it in action:

1. Invite a user to a project via teams
2. Can no longer Click Edit on the "Invited manually via teams" invite type!

#Test Planning

#Unpublished Engagement Phases No Longer Show "In Test"

What we fixed: Creating an engagement phase for the current date and unpublishing it was causing the project to show up as "In test" on the community home page. This could confuse testers since there was no active phase to participate in.

See it in action:

1. Create an engagement phase dated for today but leave it unpublished
2. Navigate to the community home page
3. The project no longer incorrectly displays as "In test"

Good to know: This also applies to Onboarding and Closure phase messaging on the community homepage.

#Miscellaneous

#WYSIWYG "Code" Style Formatting Now Matches Preview

What we fixed: The "Code" style in the WYSIWYG editor was looking great while you were editing, but then rendered completely differently once published. Multi-line code blocks were especially affected.

See it in action:

1. Open any WYSIWYG editor field
2. Type some multi-line text and apply the "Code" style
3. Save and view the published result, it now matches what you saw in the editor

#Content View Counters Now Track Homepage Views

What we fixed: Content items set to "Show on Home Page" were rendering just fine on the Project Home Page, but their view counters stubbornly stayed at zero. Views were only being counted when someone clicked through to the individual content page. Now, views are tracked for content displayed in both the main content list and the "What's New" section on the homepage.

See it in action:

1. Set a content item to "Show on Home Page"
2. Visit the Project Home Page as a tester (the content renders inline)
3. Check the view count — it now increments! Refreshing the page within the same session won't double-count.

Good to know: A view is a view regardless of where it was seen. If someone views content on the homepage and then clicks into the full page, it only counts once per session.

We've been hard at work on a major overhaul of AI-powered feedback tools, and this month they're officially out of GoEarly and into production for everyone. And there's already something new blooming in the GoEarly pipeline: AI Feedback Submission is heading into early access, giving community members a whole new way to share what they're thinking. Want in early? Get signed up now!

#Reimagined Feedback View

The feedback list has a new look, and it's more than cosmetic. We've overhauled the Feedback View to make navigating and acting on feedback faster, with a cleaner layout and a more consistent experience across the board. Whether you're triaging a full queue or digging into a single ticket, you'll find the controls where you expect them and the information you need front and center.

#AI Feedback Summary and AI Discussion Summary

Ted now reads your tickets so you don't have to start from scratch. Two new AI Summary fields are available in Feedback: one that condenses the full ticket into a quick overview, and one that summarizes the entire comment thread separately. Walk into any piece of feedback already up to speed, even if the discussion has been running for weeks.

Both summaries live right on the feedback card and are accessible from two places: expand any card in the feedback list and click the Toggle AI summary button to see them side by side, or open an individual ticket and expand the card just below the header. The Discussion Summary is context-aware, too. It takes the Feedback Summary as input, so it's grounded in the original submission rather than treating the replies in isolation. And when a new reply comes in, the Discussion Summary regenerates automatically.

Admins can enable AI Feedback Summaries per feedback type under Management > Project Configuration > Feedback Types > Modify > Feedback Properties, with the option to add custom instructions per type. For example, you might direct the AI to focus on reproduction steps for issues or emphasize sentiment for ideas. Both summary fields are also available as view and filter columns, with full support in reports, exports, integrations, and email templates as dynamic tags. Check out the AI Feedback Summaries Overview for full setup details.

#AI Reply Generator

Responding to feedback just got a lot faster. When you change the status on a feedback ticket, the AI Reply Generator automatically drafts a comment for you based on the ticket content and the status you selected. A status change to Need More Info produces a different reply than one to Fixed; the tone and content are tailored to the intent of the update.

The draft appears right in the reply field on the feedback card, accessible from both the feedback list and the individual ticket view. Review it, edit it however you like, and click Update status to post the reply and status change together. You can also clear the field entirely and submit the status update without a reply. Nothing goes out until you say so.

AI Smart Reply Drafts are enabled per feedback type under Management > Project Configuration > Feedback Types > Modify > Feedback Collection and Collaboration. Admins can also configure per-type prompt enhancements to tune the tone of generated replies, whether that's more formal for enterprise programs or more conversational for consumer ones. For full setup details, see the AI Smart Reply Drafts Overview.

#Redesigned Community AI Settings

The Community AI settings page has been rebuilt from the ground up. The new design is cleaner and easier to navigate, and it comes with brand-new licensing controls that give admins more direct control over how AI features roll out across their community. If you've been wanting a clearer picture of your AI configuration, this is the page you'll want to revisit.

#GoEarly Features

#AI Feedback Submission and AI Element Enhancement

We've got two big ones headed into GoEarly this month, and both are aimed at making feedback better before it even lands in your queue.

AI Feedback Submission gives community members a new way to share what they're experiencing: they describe it in plain language using the AI Submit button in the project navbar, and AI organizes that input into one or more structured submissions automatically. Members review everything on a provisional screen before anything is finalized, with five minutes to edit, reassign, or discard. There's also a built-in duplicate check that surfaces similar existing feedback so members can vote instead of filing a new entry. On the admin side, a Show original option lets you pull up the member's unedited input alongside the structured result.

AI Element Enhancement works at the form level. Admins can mark any feedback form field as AI-generated, so after a member submits, AI automatically populates it based on the rest of the submission. A vague title becomes a clean one. A summary field fills itself in. Admins write the prompt and choose which fields feed into it; members just fill out the same form they always have.

Want early access to both? Get signed up for GoEarly now!

#Data Engine & Custom Reports

#File Exports No Longer Tripped Up by Certain File Names

What we fixed: Exporting files from a feedback type could silently produce an incomplete zip if any uploaded files had certain reserved names (like `com1.png` or `aux.txt`). The export would show the right file count in the preview, but the zip would come up short. Those files are now handled properly during export.

See it in action:

1. Go to a feedback type that has file attachments
2. Run an export that includes files
3. Check your zip, all files should be present and accounted for, regardless of their names

#Onboarding/Recruiting

#API Import Invitations Now Accept Custom Messages As Intended

What we fixed: When setting up an API import invitation and trying to add a custom message, the form would incorrectly throw a "Write a custom message" validation error, even though you just did! Custom messages now save as expected.

See it in action:

1. Navigate to a project's Recruiting section
2. Create or edit an API import invitation type
3. Enter a custom message and submit, it'll work without a fuss!

#User Interface & Experience

#Copy Tester-Facing Links from Content and Releases

What we fixed: Well, this one's brand new! Admins can now copy a tester-facing link directly from Content and Releases admin list views — no more manually editing URLs to share with testers.

See it in action:

1. Navigate to a Content or Releases admin list view
2. Click the clipboard icon on any row
3. You'll see a "Link copied!" toast — paste it wherever you need to share it with testers

Good to know: If the short link can't be created for any reason, the full tester-facing URL is copied instead, so you're always covered.