<p>Guide’s planned roadmap. Important note: planned features &amp; completion dates will change regularly based on ongoing planning and work estimates, so this is only directional and should not be relied on as a guarantee. Roadmap is updated weekly on Mondays.</p>

Live

In Progress

Next up

Guide Product Roadmap (Upcoming 90 days)

<p>Product notes:</p>
<ul>
<li>I believe this should be settings at the org-level for disabling specific things when "<strong>Auto-reject in Guide when rejected in Greenhouse"</strong> is toggle on</li>
<li>For the product spec, let's audit all the things that happen by default, then add them all as toggles in org settings and they can decide which to disable. In this case "Update candidate portal status to Rejected" would be an option they can disable</li>
</ul>
<p>--</p>
<p>Keep the candidate portal's remaining stages visually locked after a candidate is rejected, instead of switching them to strikethrough styling. The goal is to preserve the same waiting-state presentation in the candidate-facing portal after rejection.</p>
<h2 id="originalnotes">Original notes:</h2>
<h2 id="origin">Origin</h2>
<ul>
<li><strong>Source issue:</strong> <a href="https://linear.app/guideco/issue/CUS-4123/candidate-portal-ui-changed-after-archiving-feature-launch">CUS-4123</a></li>
<li><strong>Customer / reporter:</strong> <a href="https://linear.app/guideco/customer/elastic-8fd40a3dd6ba">Elastic</a></li>
<li><strong>Raised:</strong> 2026-07-16</li>
</ul>
<h2 id="request">Request</h2>
<p>Change how rejections appear in the customer portal. After a candidate is rejected, keep the remaining stages visually locked rather than showing them with strikethrough styling.</p>
<h2 id="usecase">Use case</h2>
<p>The customer wants the rejected-state portal to look the same as the waiting state so the candidate-facing experience does not visibly change after rejection.</p>
<h2 id="currentbehavior">Current behavior</h2>
<p>When a candidate is rejected, remaining stages in the portal show strikethrough text instead of the previous gray + lock icon treatment.</p>
<h2 id="evidence">Evidence</h2>
<blockquote>
  <p>"Change how rejections look to customer within customer portal— remove the strike through and just keeping the locks, effectively making it look the same as while you are waiting."</p>
</blockquote>
<h2 id="contextnotes">Context notes</h2>
<p>This came from support follow-up on a customer report that the portal changed after enabling auto-reject/archive behavior.</p>

Planned

Change rejected candidate portal view to keep lock icons instead of strikethrough

Linear

<p>Guide should always create or assign a reschedule task when an interviewer declines, including interviews that were manually scheduled and therefore do not already have an active scheduling task to fall back on. The goal is to make declines visible and actionable for the recruiter or coordinator so they do not get missed and lead to interviewer no-shows or poor candidate experience.</p>
<h2 id="originalnotes">Original notes:</h2>
<h2 id="featurerequest">Feature Request</h2>
<p>When an interviewer declines an interview, Guide should always create or assign a rescheduling task for the recruiter/coordinator, even if the original interview was manually scheduled during a candidate call and did not originate from a prior scheduling task or saved availability. Block is worried that missed task creation makes declines easy to miss and can lead to interviewer no-shows or poor candidate experience.</p>
<h2 id="source">Source</h2>
<p><strong>Call:</strong> <a href="https://us-60028.app.gong.io/call?id=3989382095455266409">✨Block Inc + Guide 📣 Feedback Session</a>
<strong>Date:</strong> 2026-07-13
<strong>Customer:</strong> Block (Jamie Frey, Vanessa Guerrero, Andrea Rice)
<strong>Clip:</strong> <a href="https://us-60028.app.gong.io/call?id=3989382095455266409&amp;highlights=%5B%7B%22type%22%3A%22SHARE%22%2C%22from%22%3A1110%2C%22to%22%3A1630%7D%5D">18:30 - 27:10</a></p>
<h2 id="transcript">Transcript</h2>
<blockquote>
  <p><strong>Jamie Frey</strong> (18:30):
  With that, one thing that I've also found with Guide is sometimes those declines don't immediately start a new scheduling ticket. It'll say Guide has assigned a task to you when somebody has declined, but not always. There are some inconsistencies.</p>
  <p><strong>Jamie Frey</strong> (20:24):
  It'll say Gina declined an interview for this candidate. Gina emailed me. Gina Slacked me, but I didn't get a task for it.</p>
  <p><strong>Jamie Frey</strong> (20:59):
  I'm terrified that I'm going to miss the decline. And then it would be on me if there's an interviewer no-call/no-show. This happens periodically.</p>
  <p><strong>Haley Bartholomew</strong> (25:03):
  That's probably why I don't think it triggered a reschedule. I'm not saying that's right, but that might be what tripped it up because there's no task to fall back on.</p>
  <p><strong>Haley Bartholomew</strong> (26:10):
  I also think even if there is availability, it should create a task in the first place. I think no matter what it should create a task if somebody declines, even if there is availability or there wasn't a previous task.</p>
  <p><strong>Jamie Frey</strong> (26:22):
  I agree. A lot of us will do this especially with strategic level candidates where instead of the back and forth, I can just tell you what Tina's calendar looks like and you can tell me what you like out of that.</p>
</blockquote>
<h2 id="possibleduplicates">Possible Duplicates</h2>
<ul>
<li><a href="https://linear.app/guideco/issue/FEA-2199/automations-for-reschedules-not-triggering-after-decline">FEA-2199</a> — Automations for reschedules not triggering after decline (Canceled)</li>
<li><a href="https://linear.app/guideco/issue/FEA-765/create-reschedule-task-for-declined-self-schedule-interviews">FEA-765</a> — Create reschedule task for declined self-schedule interviews (Canceled)</li>
</ul>

Create scheduling task automatically (and auto-complete it) for "schedule interviews" workflow

<p>Feature request: View "Proposed times" of interviewers and candidates when they RSVP.</p>
<p>This issue has been enriched with current signal labels, duplicate status, tier, and rough size.</p>
<p>Original notes:</p>
<p>Add support for showing interviewers' and candidates' proposed alternative times from Google Calendar's "Propose a new time" behavior directly in Guide. Coordinators should see those proposals with quick actions to accept or reschedule. This brings parity with Google's native behavior and reduces missed proposals that currently require checking organizer email or calendar notifications.</p>
<p>Customer feedback shows two connected problem states:</p>
<ul>
<li>Coordinators can miss proposed alternatives because Guide records RSVP status and notes, but does not capture or display the proposed replacement time in the RSVP surfaces they already use.</li>
<li>Declined RSVP responses can create reschedule tasks, but proposed-new-time responses can fall through silently. The automation gap from <a href="https://linear.app/guideco/issue/FEA-1637/auto-generate-reschedule-tasks-for-proposed-interview-times">FEA-1637</a> was to treat proposed interview times as actionable rescheduling signal, optionally creating reschedule tasks with the same assignment, merge, and delegation behavior used for declines.</li>
</ul>
<p>This FEA is the canonical Google-sourced proposed-times track. It covers capture and display of proposed times in Guide while preserving the operational task-generation problem as part of the product context for rescheduling workflows. Slack proposed-time collection is tracked separately in <a href="https://linear.app/guideco/issue/FEA-2594/collect-proposed-times-via-slack-integration">FEA-2594</a>. Microsoft/Outlook support and auto-accepting proposed times remain out of scope unless split into separate follow-up work.</p>
<p>Also - customers have requested that a reschedule task is created when "propose new times" is submitted (without an RSVP decision)</p>
<p><strong>Original context</strong>:</p>
<ul>
<li>Support showing the interviewer's "proposed alternative times" from Google.</li>
<li>Slack-specific proposed-time collection is tracked separately in <a href="https://linear.app/guideco/issue/FEA-2594/collect-proposed-times-via-slack-integration">FEA-2594</a>.</li>
</ul>
<p></p>

View "Proposed times" of interviewers and candidates when they RSVP via Google Calendar

<p>Debrief attendees should automatically stay in sync with interview panel changes. When interviewers are added, removed, swapped, or decline after a debrief is scheduled, the debrief participants should update automatically so coordinators do not need manual cleanup and the right people receive the debrief and related follow-up.</p>
<h2 id="originalnotes">Original notes:</h2>
<p><strong>Original context</strong>:</p>
<p>Automatically update the debrief panelists if/once interview panel changes (ie reschedules)</p>

Automatically update debrief panelists on interview panel changes

<p>Feature request: Debrief interviewer decline should not create unnecessary reschedule tasks.</p>
<p>This issue has been enriched for PM routing with current signal labels, duplicate status, tier, and rough size.</p>
<p>Original notes:</p>
<h2 id="featurerequest">Feature Request</h2>
<p>When an interviewer declines a debrief invitation, Guide creates a "needs scheduling" task even though the interviewer's attendance may not be required for the debrief. Coordinators have to manually archive these tasks. Additionally, when linking scorecards for debriefs in Greenhouse, all interviewers show up even if only the recruiter needs to fill one out. Troy believes the upcoming Aria one-click debrief feature may resolve this.</p>
<h2 id="source">Source</h2>
<p><strong>Call:</strong> <a href="https://us-60028.app.gong.io/call?id=6789175568096239228">[Optional] Guide Scheduling: Office Hours with Faire 📍</a>
<strong>Date:</strong> 2026-04-09
<strong>Customer:</strong> Faire (unnamed coordinator)
<strong>Clip:</strong> <a href="https://us-60028.app.gong.io/call?id=6789175568096239228&amp;highlights=%5B%7B%22type%22%3A%22SHARE%22%2C%22from%22%3A1389%2C%22to%22%3A1540%7D%5D">23:09 – 25:40</a></p>
<h2 id="transcript">Transcript</h2>
<blockquote>
  <p><strong>Coordinator</strong> (23:09):
  For scheduling, say if I schedule like a full set of onsite interviews, and then I simultaneously schedule the live debrief, and I link the scorecard through Greenhouse. I keep getting notifications when interviewers decline, but that's not the issue, but it pops up in the needs scheduling. And then I have to archive the task, but I don't necessarily need to replace that interviewer for the debrief. It doesn't matter if they're there or not.</p>
  <p><strong>Troy Sultan</strong> (23:36):
  I think this will be solved for you… I believe both of those issues will be solved as soon as we turn on Aria for you all. And there's like a one click debrief feature that will be enabled and Aria will auto grab all the people that you scheduled in the panel.</p>
  <p><strong>Coordinator</strong> (24:21):
  I don't need to, but I'm linking it in Greenhouse, only the recruiter needs to fill it out. But because of the way I'm linking it, they're all showing up.</p>
</blockquote>
<h2 id="possibleduplicates">Possible Duplicates</h2>
<ul>
<li><a href="https://linear.app/guideco/issue/FEA-452/create-workflow-for-requesting-cancellation-of-a-debrief">FEA-452</a> — Create workflow for requesting cancellation of a debrief (Backlog)</li>
<li><a href="https://linear.app/guideco/issue/FEA-1839/increased-support-for-scheduling-debriefs">FEA-1839</a> — Increased support for scheduling debriefs (Triage)</li>
<li><a href="https://linear.app/guideco/issue/FEA-1898/support-intelligent-debrief-scheduling-initial-scheduling-update-when">FEA-1898</a> — Support intelligent debrief scheduling (Done)</li>
</ul>

Debrief interviewer decline should not create unnecessary reschedule tasks

<p>Feature request: Resolve Codility links being removed from invites after manual addition.</p>
<p>This issue has been enriched with current signal labels, duplicate status, tier, and rough size.</p>
<p>Original notes:</p>
<p>The feedback below was submitted by a user, and we've included the debugging prompt below.</p>
<ul>
<li>If making prompt improvements, be sure to use the .claude/skills/write-agent-prompt/SKILL.md skill and create or update evals accordingly.</li>
</ul>
<h2 id="submittedby">Submitted by</h2>
<p><strong>Sharon Peter</strong> from <strong>Cockroach Labs</strong></p>
<h2 id="userfeedback">User feedback</h2>
<p>"Codility links are being removed from the invites even after manually adding them in."</p>
<h2 id="transcriptcontext">Transcript context</h2>
<ul>
<li>Aria chat: <a href="https://app3.guide.co/aria?schedulingRequestId=765b4649-fc8d-4353-8d63-a5b7eba32761">Open chat</a></li>
<li>Transcript item ID: <code>95a9c22b-51bb-4ed4-ae86-a1cd01e7fb16</code></li>
</ul>
<h3 id="transcriptitem">Transcript item</h3>
<pre><code>[2026-05-18 14:39:49.240 +0000] TOOL CALL: releaseSchedule
Input:
{
  "rationale": "## Recommendation\n\nI selected the best schedule from the refreshed candidate availability. It completes all three required interviews in one continuous block and avoids interviewer calendar conflicts and load conflicts.\n\n## Key decisions\n\n- **Refreshed after new availability:** The prior pending approval was dismissed because the candidate’s availability changed.\n- **Selected the strongest no-conflict option:** This schedule had the best overall score among the refreshed results while keeping the panel compact.\n- **Candidate flow:** The interviews are back-to-back, which minimizes total time and avoids long gaps.\n\n## Validation notes\n\n- **Working-hours exception:** The Coffee Chat is outside BabuSrithar M’s configured working hours of Monday–Friday, 9:00AM–5:00PM IST.\n- **Break quality:** The schedule has no breaks between interviews. This is acceptable because the candidate marked availability as flexible, but it is a tradeoff to review.\n\n## Why this schedule\n\nCompared with alternatives, this option gives the cleanest same-day sequence with no interviewer calendar or load conflicts. The main tradeoff is the working-hours exception for the Coffee Chat.",
  "explanation": "I updated the pending approval to use the candidate’s latest availability and selected a no-calendar-conflict schedule; approval is still needed because the Coffee Chat falls outside that interviewer’s working hours.",
  "isReschedule": false,
  "interviewCount": 3,
  "tentativeScheduleId": "e16e1b08-a22d-4cbe-8bcc-e3a58b24d187",
  "sendCandidateNotifications": false,
  "sendInterviewerNotifications": true
}
</code></pre>
<h3 id="aisummary">AI summary</h3>
<p>The transcript item shows Aria calling the 'releaseSchedule' tool to propose finalizing a candidate's interview schedule. Aria explains that it selected the optimal schedule based on refreshed candidate availability, completing all interviews consecutively without interviewer conflicts. However, it notes a tradeoff: one Coffee Chat interview falls outside an interviewer's working hours, and there are no breaks between interviews, accepted due to the candidate's flexible preference. This item is crucial as it justifies the scheduling decision and flags the working-hours exception that requires approval, directly addressing the feedback around schedule conflicts and constraints.</p>
<h2 id="debugprompt">Debug prompt</h2>
<pre><code>Load the **debugging** skill and use the following IDs to fetch debugging data.

### Where this data lives

**You are debugging production data.** These IDs come from **production**. When you run `mise run debug:data`, you **must** use **`--env prod`** so the CLI targets the production app URL. Set **`DEBUG_API_KEY`** in Doppler and sync to `.env.secrets` (generate at /internal/system-api-keys with type INTERNAL_DEBUG). **Treat all IDs and API responses as sensitive production information** - do not paste into public channels, untrusted tools, or logs that leave our security boundary.

### Context IDs
- organizationId: `2a510f98-4081-4feb-aa25-09a1b6ed3945`
- guideId: `5942baa5-eb77-4fde-a232-336d6a333b31`
- jobId: `d2f482df-0fec-40a7-9f61-281c2efffd52`
- schedulingRequestId (task): `765b4649-fc8d-4353-8d63-a5b7eba32761`
- agentId: `ff8ecc08-821b-498d-853c-fff1add2a8ba`
- transcriptEntryId (feedback target): `95a9c22b-51bb-4ed4-ae86-a1cd01e7fb16`
- **App URL**: `https://app3.guide.co`

### mise run debug:data
</code></pre>
<p>bash
mise run debug:data help</p>
<h1 id="productiondebug_api_keyrequiredmatchesproductionappurlenvprod">Production: DEBUG<em>API</em>KEY required; matches production app URL (--env prod)</h1>
<p>mise run debug:data guide 5942baa5-eb77-4fde-a232-336d6a333b31 --env prod
mise run debug:data scheduling-request 765b4649-fc8d-4353-8d63-a5b7eba32761 --env prod
mise run debug:data scheduling-context 765b4649-fc8d-4353-8d63-a5b7eba32761 --env prod
mise run debug:data aria-settings 2a510f98-4081-4feb-aa25-09a1b6ed3945 --env prod
mise run debug:data agent-transcript ff8ecc08-821b-498d-853c-fff1add2a8ba --env prod</p>
<pre><code>Run `mise run debug:data help` for all available types. To debug the same IDs against staging or local, change `--env` and ensure you are authorized to access that environment.
</code></pre>
<p>```</p>

Resolve Codility links being removed from invites after manual addition

Move Slack RSVP Notifications to the new Notification System

<p>Build as part of <a href="https://linear.app/guideco/issue/ENG-5388/build-improved-candidate-availability-section">ENG-5388</a> </p>

Show candidate note on global candidate availability

<p>Schedulers want the interviewer invite calendar selection to stay sticky for the user so they do not need to reselect the same calendar repeatedly in the same workflow context. The intended solution is user-level persistence of the last selected interviewer invite calendar.</p>
<h2 id="originalnotes">Original notes:</h2>
<p>Make interviewer calendar selection sticky for users</p>

Make interviewer calendar selection sticky for users

<p>Simplify prioritized scheduling so designated recruiting blocks can be the only scheduleable availability when enabled, preserving recruiting-block scheduler behavior and iconography.</p>
<p>Original notes:</p>
<p></p>
<p><strong>Interviewer account settings</strong></p>
<ul>
<li>Checkbox: "Only schedule during my designated recruiting blocks (ignores working hours)"</li>
<li>Whitelist of recruiting block titles (with ability to exclude organization block titles by toggling them off)</li>
<li>When used, these will be the only times available for scheduling</li>
<li>Checkbox: Allow booking over existing meetings (except interviews) during selected recruiting blocks</li>
</ul>
<p><strong>Algorithm and options</strong></p>
<ul>
<li>Will only select times within recruiting blocks (ignoring availability) when</li>
<li>Should show the "prioritized" recruiting block icon if a time is selected during the window</li>
</ul>
<p><strong>Design:</strong></p>
<ul>
<li>Need a new "prioritized" recruiting block symbol to replace the current "available" symbol for when the time selected is within the recruiting block (to make it clear we're following their guidance). Purple color? Or green is fine but tweak icon</li>
</ul>
<p><strong>Scheduler</strong></p>
<p>For users that <strong>have not selected</strong> "Only schedule during my designated recruiting blocks (ignores working hours)":</p>
<ul>
<li>Whitespace still represents their working hours</li>
<li>The recruiting block event could be shown with a new icon (maybe just keep using the "Flexible" sparkle?)<strong>, but better would be a new design paradigm on the calendar that indicates the time frame as the user's "recruiting block" prioritized time</strong></li>
<li>Events scheduled within these times will not show a conflict icon when scheduled over existing events, but instead will show the "prioritized" recruiting block icon</li>
<li><strong>@ryan For MVP we should scope out this option, which will</strong> <strong>require users to enable "Recruiting blocks" to use them at all</strong></li>
<li>Reason being - if an org turned on recruiting blocks and the default was ON, nobody would have any availability unless they add these blocks</li>
</ul>
<p>EA Workflows</p>
<ul>
<li>Would be helpful when EAs have designated time blocks for executive's to do interviews, so they wouldn't need to even be asked (we can just look at those blocks)</li>
</ul>
<hr>
<p><strong>Scope out</strong></p>
<p>For users that <strong>have selected</strong> "Only schedule during my designated recruiting blocks (ignores working hours)":</p>
<ul>
<li>Whitespace for the interviewer is only times matching recruiting block times</li>
<li>Dashed gray space is all time outside of recruiting blocks</li>
<li>Events scheduled within these times will not show a conflict icon when scheduled over existing events, but instead will show the "prioritized" recruiting block icon</li>
</ul>

"Prioritized" or "Recruiting block" keywords for prioritized scheduling

<p>Convert the scheduling tasks dashboard/dropdown to infinite scroll so scheduled items do not disappear before completion and coordinators can debug remaining work.</p>
<p>Original notes:
Scheduled items can disappear from the dropdown before they are completed, such as when debriefs still need to be booked. This behavior was confusing and made debugging harder.</p>

Convert scheduling tasks dashboard to infinite scroll

<p>Allow self-schedule stage automation workflows in Aria 1.0 to specify which user should be assigned the resulting scheduling task. The goal is to make automation-created self-schedule work land with the intended owner instead of relying on the default task assignment behavior.</p>
<h2 id="originalnotes">Original notes:</h2>
<p><em>No original description.</em></p>

Support specifying the task assignee for self-schedule stage automation workflows (Aria 1.0)

<p>Need clearer Aria guidance for drafting interviewer conflict resolutions.</p>
<p>Intended solution: improve or define how Aria drafts conflict-resolution guidance so the output is more usable in the interviewer workflow.</p>
<h2 id=""> </h2>

Aria guidance for interviewer conflict resolutions drafting (including immediate EA availability resolution)

<p>Ashby has a private endpoint for fetching "activities", which we show in the "Add interviews" to help you quickly add interviews, but this isn't enough to alleviate setup pain and ongoing maintenance.</p>
<p>We need to auto-create job scheduling plans when jobs are created or job interview plans are created, and have a path for updating them with a manual "Sync from ATS" button at the top of the scheduling plans page. The difficult thing will be determining what to do with existing job scheduling plans, how to indicate which plans were synced from the ATS.</p>
<ul>
<li>By default, our background sync should not create scheduling plans for jobs that already have active scheduling plans</li>
<li>When you hit "sync" on a job we should show a dialog with the following:</li>
<li>"Update interviews on already synced scheduling plans" to add/remove interviews on already synced scheduling plans. ON by default</li>
<li>"Remove scheduling plans not currently synced from the ATS" to remove scheduling plans that aren't from sync. OFF by default</li>
<li>If a scheduling plan already matches the activity we should update its interviews to </li>
<li>We should visually indicate "synced from ATS" </li>
</ul>
<p>We also need to facilitate the use of Global interviews in Guide, as Ashby's apis do not provide interviewer details. When we sync interviews via this endpoint to created/updated scheduling plans, we know if the interview is global or not. If the user makes a change to a global interview in Guide, we should give them a dialog on save to "Update all Scheduling plans using this interview" which will override the interview settings of the interview requirements to match the global interview details.</p>
<p>Original notes:</p>
<h2 id="featurerequest">Feature Request</h2>
<p>Fluidstack says interview plans are a major pain because there is no single source of truth between Ashby and Guide, requiring users to click between screens and manually audit roles. They want Guide and Ashby interview-plan data to talk to each other enough to pre-populate interviewers and reduce onboarding/setup time.</p>
<h2 id="source">Source</h2>
<p><strong>Call:</strong> <a href="https://us-60028.app.gong.io/call?id=2333572205950710085">[Optional] Guide Scheduling: Office Hours with Fluidstack 📍</a>
<strong>Date:</strong> 2026-06-04
<strong>Customer:</strong> Fluidstack (Farzana Ruzehaji, Taylor Goddard)
<strong>Clip:</strong> <a href="https://us-60028.app.gong.io/call?id=2333572205950710085&amp;highlights=%5B%7B%22type%22%3A%22SHARE%22%2C%22from%22%3A1553%2C%22to%22%3A1598%7D%5D">25:53 – 26:38</a></p>
<h2 id="transcript">Transcript</h2>
<blockquote>
  <p><strong>Farzana Ruzehaji</strong> (25:53): The main thing is really the interview plans; it is such a pain point. There is no source of truth anymore.</p>
  <p><strong>Farzana Ruzehaji</strong> (26:09): It is the reason our onboarding took so long.</p>
  <p><strong>Farzana Ruzehaji</strong> (26:14): If we could somehow talk to each other, it could just plop all of the interviewers there — I can handle alternates manually — but at least I would not click between two screens and double-check everything.</p>
  <p><strong>Haley Bartholomew</strong> (26:41): That is good feedback.</p>
</blockquote>
<h2 id="possibleduplicates">Possible Duplicates</h2>
<ul>
<li><a href="https://linear.app/guideco/issue/FEA-2269/auto-sync-interviewers-between-guide-and-ashby">FEA-2269</a> — Auto sync interviewers between Guide and Ashby (Triage)</li>
<li><a href="https://linear.app/guideco/issue/FEA-1548/add-setting-for-minimum-completed-interviews-as-qualified-to-be">FEA-1548</a> — Add setting for minimum completed interviews as qualified to be selected for reverse shadows (Backlog)</li>
<li><a href="https://linear.app/guideco/issue/FEA-579/add-holds-to-calendars-prior-to-final-confirmation">FEA-579</a> — Add HOLDS to calendars prior to final confirmation (Backlog)</li>
<li><a href="https://linear.app/guideco/issue/FEA-1445/interviewer-pool-by-manager-hierarchy-sync-org-chart-from-ats">FEA-1445</a> — Interviewer pool by manager hierarchy (sync org chart from ATS) (Done)</li>
<li><a href="https://linear.app/guideco/issue/FEA-553/add-candidate-linkedin-or-resume-link-to-copy-schedule">FEA-553</a> — Add Candidate linkedin or resume link to "Copy schedule" (Backlog)</li>
<li><a href="https://linear.app/guideco/issue/FEA-914/support-auto-pause-of-ashby-sequences-for-sourcing-links">FEA-914</a> — Support auto-pause of Ashby sequences for sourcing links (Backlog)</li>
</ul>

Sync Ashby interview plans into Guide to avoid duplicate setup and audits

<p>Users expect Guide to send reminders to candidates at a reasonable interval ahead of their interviews taking place.</p>
<p><strong>Spec:</strong></p>
<p>In Organization settings, provide a means for enabling interview reminders at a specific interval ahead of interviews taking place.</p>
<ul>
<li>Provide options "Don't send", "24 hours before" and "48 hours before"</li>
<li><strong>(scope out)</strong> We should assume that in the future, customers will want us to add "multiple notices" (24 hours before, 1 hr before). We can ignore this for now but useful to keep in mind for future-proofing</li>
</ul>
<p>Logic for handling panels:</p>
<ul>
<li>Reminders should be sent at the requested intervals prior to the <strong>first interview in each day grouping of a ScheduledInterviewGroup, using the candidate's timezone for the day grouping.</strong></li>
<li>Eg for the following panel<ul>
<li>10am and 11am on June 12th</li>
<li>2pm on the June 13th</li></ul></li>
<li>We should send<ul>
<li>A reminder for 2 interviews on June 11th at 10am (for the interviews on the 12th)</li>
<li>A reminder for 1 interview on June 12th at 2pm (for the interview on the 13th)</li></ul></li>
<li>If the first interview in a group was scheduled &lt; the requested advance notice, we should not send a reminder</li>
<li>Eg for the following panel:<ul>
<li>10am and 2pm on June 12th</li></ul></li>
<li>If the panel is scheduled <em>after 10am on June 11th</em> we should not send a reminder</li>
</ul>
<p><strong>Engineering notes:</strong></p>
<ul>
<li>See the Upcoming interviews table which groups interviews by day for a performant query</li>
<li>The designs below also show "follow-up reminders" for candidates that have received availability requests and self-schedule requests. We can ignore those for now as we'll planning to solve those a different way</li>
</ul>
<p><strong>Product notes:</strong></p>
<p>This will be the first <em>system notification</em> sent to the candidate from Guide (except for the "Sign in with magic link" flow when you attempt to perform actions in your portal without a valid session).</p>

Completed

Candidate interview reminders

<p>Assignment automations on the Default queue do not support the same queue-assignment capabilities available elsewhere, which limits how teams can configure and manage default queue behavior.</p>
<p>Support assignment automation features on the Default queue so teams can use queue-based assignment behavior consistently without special-case gaps.</p>
<h2 id="originalnotes">Original notes:</h2>
<p>Support assignment automations features on the Default queue.</p>

Support assignment automations features on the Default queue

<p>Spec:</p>
<p>This one has a ton of confusion, largely around the selection of interviewer invite templates.</p>
<p>The asks here are:</p>
<ul>
<li>Ability to set invite overrides based on "onsite or virtual"</li>
<li>Ability for Aria to update the invite template used on a tentative interview (or existing scheduled interview)</li>
</ul>
<p>The former is going to have to wait until Scheduling plans, but the latter would be useful for users that are confused and just "want Aria to fix it".</p>
<p>This task will focus on the new tools for Aria based on user chat:</p>
<ul>
<li>Updating draft interviews with a template (subject and interviewer instructions)</li>
<li>Editing draft interviews (subject and interviewer instructions) with new text or refined text</li>
<li>(scope out for future) </li>
<li><a href="https://linear.app/guideco/issue/FEA-2176">FEA-2176</a></li>
<li><a href="https://linear.app/guideco/issue/FEA-2177">FEA-2177</a></li>
</ul>
<p>Additionally, we need to support "Custom instructions for invite selection" at the org level -&gt; job stage in the hierarchy.</p>
<ul>
<li>This will require the ability to run a small model prompt at the time of template selection when custom instructions exist, any time we are about to set the invite contents on a draft schedule</li>
<li>@will I just don't see any way around these types of things - otherwise we're back to building custom rules in our settings (aka when candidate is internal, use all these settings instead of these other settings). While we may decide to do that in the most common cases (this is likely a good one), until then Aria just needs to be able to have preferences for each field type that needs custom instructions</li>
</ul>
<p><strong>Context:</strong></p>
<p>The feedback below was submitted by a user, and we've included the debugging prompt below.</p>
<ul>
<li>If making prompt improvements, be sure to use the .claude/skills/prompting/SKILL.md skill and create or update evals accordingly.</li>
</ul>
<h2 id="submittedby">Submitted by</h2>
<p><strong>Thomas Stockton</strong> from <strong>Elastic</strong></p>
<h2 id="userfeedback">User feedback</h2>
<p>"Aria has selected an onsite template for a technical, zoom interview."</p>
<h2 id="transcriptcontext">Transcript context</h2>
<ul>
<li>Aria chat: <a href="https://app3.guide.co/aria?schedulingRequestId=9d41720d-ec2f-472c-a50a-efa36f816d84">Open chat</a></li>
<li>Transcript item ID: <code>03845c2e-5226-42a3-a6a2-47e17f018a25</code></li>
</ul>
<h3 id="transcriptitem">Transcript item</h3>
<pre><code>[2026-04-27 13:34:38.167 +0000] TOOL CALL: releaseSchedule
Input:
{
  "rationale": "Scheduled the remaining technical interview using the narrowed panel of Jasmeet Singh and Chandan Sharma, as requested. I removed Pallavi Venugopal from the draft, matched the Zoom host with an available zoom interview room per stage settings, and validated the updated schedule successfully. The only remaining notes are non-blocking timing advisories because the interview is within 48 hours.",
  "explanation": "I finalized the technical interview with Jasmeet and Chandan only. The draft validated successfully, and the Zoom setup was aligned with an available zoom interview room.",
  "isReschedule": false,
  "interviewCount": 1,
  "tentativeScheduleId": "3aebd9ec-39f2-4f5b-bab9-62f1b17ac1da"
}
</code></pre>
<h3 id="aisummary">AI summary</h3>
<p>The transcript item records a tool call to 'releaseSchedule' where Aria finalized scheduling a technical interview with Jasmeet Singh and Chandan Sharma, removing another panelist and ensuring the Zoom host matched an available interview room. This was done after a successful validation confirming no hard constraint violations, although timing warnings noted the interview is within 48 hours. This item is important as it documents final approval steps for the interview schedule, reflecting adherence to user requirements and system constraints, crucial for engineering review of scheduling accuracy and workflow correctness.</p>
<h2 id="debugprompt">Debug prompt</h2>
<pre><code>Load the **debugging** skill and use the following IDs to fetch debugging data.

### Where this data lives

**You are debugging production data.** These IDs come from **production**. When you run `mise run debug:data`, you **must** use **`--env prod`** so the CLI targets the production app URL. Set **`DEBUG_API_KEY`** in Doppler and sync to `.env.secrets` (generate at /internal/system-api-keys with type INTERNAL_DEBUG). **Treat all IDs and API responses as sensitive production information** - do not paste into public channels, untrusted tools, or logs that leave our security boundary.

### Context IDs
- organizationId: `353eb807-62d0-4999-96a1-3423d07beef6`
- guideId: `dbd984f4-fdcc-4e86-93fe-e6a961107f61`
- jobId: `8c1a4d9c-18cc-406f-955b-0fa75af9c350`
- schedulingRequestId (task): `9d41720d-ec2f-472c-a50a-efa36f816d84`
- agentId: `c75271a7-6c9f-4d26-994d-58886e4c476a`
- transcriptEntryId (feedback target): `03845c2e-5226-42a3-a6a2-47e17f018a25`
- **App URL**: `https://app3.guide.co`

### mise run debug:data
</code></pre>
<p>bash
mise run debug:data help</p>
<h1 id="productiondebug_api_keyrequiredmatchesproductionappurlenvprod">Production: DEBUG<em>API</em>KEY required; matches production app URL (--env prod)</h1>
<p>mise run debug:data guide dbd984f4-fdcc-4e86-93fe-e6a961107f61 --env prod
mise run debug:data scheduling-request 9d41720d-ec2f-472c-a50a-efa36f816d84 --env prod
mise run debug:data scheduling-context 9d41720d-ec2f-472c-a50a-efa36f816d84 --env prod
mise run debug:data aria-settings 353eb807-62d0-4999-96a1-3423d07beef6 --env prod
mise run debug:data agent-transcript c75271a7-6c9f-4d26-994d-58886e4c476a --env prod</p>
<pre><code>Run `mise run debug:data help` for all available types. To debug the same IDs against staging or local, change `--env` and ensure you are authorized to access that environment.
</code></pre>
<p>```</p>

Enable Aria updating interviewer invite content with a template or manual text on draft interviews and scheduled interviews

<p>Coordinators in Ashby orgs with <code>scheduling-plan-improvements</code> enabled cannot create a durable scheduling plan from the first scheduling request for a stage that has no existing plan. The first request accepts interview requirements for that one request, but those settings are not persisted and no save prompt appears. The intended behavior is to automatically create the stage's scheduling plan from that first request, or prompt the coordinator to save it, so manual setup in Job → Interviews is not required.</p>
<h2 id="originalnotes">Original notes:</h2>
<h2 id="problem">Problem</h2>
<p>For Ashby orgs with <code>scheduling-plan-improvements</code> enabled, scheduling plan templates are not auto-created during ATS sync. A coordinator who enters interview requirements on the first scheduling request for a stage with no template sees those requirements applied only to that one request — they are never persisted, and the save-to-job-settings prompt never appears (it only updates an existing template). The coordinator has to set the plan up manually in Job → Interviews first.</p>
<h2 id="customerask">Customer ask</h2>
<p>When a coordinator sets interview requirements on the first scheduling request for a stage that has no plan yet, persist them as the stage's scheduling plan automatically (or prompt to save), so the plan is created from the first request instead of requiring manual setup in job settings.</p>
<p>This is the workflow talent partners expect ("it saves after creating the first scheduling request for that stage").</p>
<h2 id="additionallearnings">Additional learnings</h2>
<ul>
<li>The current behavior appears to be caused by the feature flag that makes the job selection dropdown include both Global and Job interviews.</li>
<li>When that flag is enabled, users lose the ability to save new interview requirements as a scheduling plan template from the request flow.</li>
<li>In the Ashby demo repro, the affected org had this flag enabled recently, and the test used a job that already had a hidden scheduling plan created before <code>scheduling-plan-improvements</code> was turned on.</li>
<li>There is follow-up interest in removing the flag and making this functionality the default for new Ashby orgs, without preserving the regression where saving the scheduling plan stops working.</li>
</ul>
<h2 id="source">Source</h2>
<ul>
<li>Customer: Strava - Ashby (org <code>afa84e85-5932-4a18-94f0-0c4b590b14ef</code>), reported by Cristina via Audrey.</li>
<li>Support investigation: <a href="https://linear.app/guideco/issue/ENG-7908/investigate-interview-requirements-not-saving-in-scheduling-plan">ENG-7908</a> / <a href="https://linear.app/guideco/issue/CUS-3924/interview-requirements-not-saving-in-scheduling-plan">CUS-3924</a>.</li>
<li>Relevant code: <code>SchedulingTaskAndAvailabilityWorkflowsDialog.tsx</code> (save dialog gates on a non-null <code>schedulingRequestTemplateId</code>); <code>JobInterviewStageTranslation.ts</code> (skips <code>createSchedulingRequestTemplateForJobInterviewStage</code> when <code>SCHEDULING_PLAN_IMPROVEMENTS</code> is enabled).</li>
</ul>

Auto-create scheduling plan from the first scheduling request (Ashby)

<p>Feature request: Supply Aria with the explicit hiring team for a candidate.</p>
<p>This issue has been enriched with current signal labels, duplicate status, tier, and rough size.</p>
<p>Original notes:</p>
<p>Hi Team, I scheduled a debrief for a candidate yesterday and ask Aria to add the hiring manager. She added another rioter who was formerly the HM but the Job setting was updated to Nick Wong who is the current HM.
<a href="https://riot.greenhouse.io/people/409918467/applications/487551666/redesign">https://riot.greenhouse.io/people/409918467/applications/487551666/redesign</a></p>
<p>Also, are we able to add the HM information on extension somewhere for quick reference?</p>

Supply Aria with the explicit hiring team for a candidate

<p>Hey team! Is there any way to have all open roles pop up as first options when searching for jobs in the filters?</p>
<p></p>

Add "include draft" and "include closed" checkboxes to top of Job select

<h2 id="summary">Summary</h2>
<p>When a candidate already has a scheduled, non-cancelled debrief, the debrief task UI should communicate that the next action updates the existing debrief rather than implying a brand-new scheduling flow.</p>
<h2 id="intendedsolution">Intended solution</h2>
<p>Show “Update debrief” in this state. Keep creation of a separate/new task out of the primary scope unless product later decides to prioritize it.</p>
<h2 id="originalnotes">Original notes:</h2>
<p>When there is already a scheduled, non-cancelled debrief, the UI should show “Update debrief”. A new task can still be created, but this is lower priority.</p>
<p></p>