Leader Rounds App Overview

Leader Rounds App Overview

Leader Rounds is structured leader rounding on staff: a leader has a short, recurring conversation with each direct report — what’s working, who deserves recognition, what’s broken, what do you need — and every issue raised becomes a tracked stoplight item the person who raised it can see. Coverage rolls up the reporting line so a director knows which leaders are rounding and what’s stuck.

Find it at Apps → Leader Rounds (or /apps/leader-rounds); on a phone it opens as a mobile capture surface at /m/apps/leader-rounds.

Key Concepts (How It Works)

A round is one recorded conversation: a leader, a subject (the person rounded on), a date, typed answers to a template’s questions, and optional private notes. Rounds are records of conversations that already happened — you can’t schedule a future round, and a round dated in the future is rejected. A round can also be skipped with a required reason (for example, the person is on leave), so an untaken round is a recorded decision rather than a gap.

The two visibility tiers are the app’s most important rule. The shared tier — the answers and the issues raised — is visible to the subject: they see the rounds about them, the items they raised, current status, and the written reason whenever something is closed without a fix. The private tier — the leader’s private notes — is visible to the leader and the leader’s own management chain, and by default to every administrator of your company (a setting an admin can turn off, under Settings → Who Can Do What). It is never shown to the subject, and it is never available to the AI assistant.

The stoplight is derived, never hand-set. An issue raised in a round lives on the platform’s corrective-action register with an owner and a due date. Its color is computed: green when resolved, red when it’s past due or was closed without a fix, yellow otherwise. Closing an issue as won’t-fix requires a written reason, because that reason is shown to the person who raised it. Every status change also lands an audit note on the issue’s timeline, in the same transaction as the change.

Cadence is an obligation, not a calendar event. The app never creates meetings on its own. Instead it answers one question per direct report: have you rounded on this person in the last N days? The due list computes itself from the org chart, so transfers and new hires are picked up automatically and deactivated people never appear.

Rounding Programs decide which conversation, and on what clock. A program is an ordered rule: an audience (job title, department, location), optionally a leader filter (the leader’s job title or department, or specific leaders chosen by name), a template, and a cadence. For each leader-and-person pair the first program that matches wins, and it supplies both the template the capture form opens with and the cadence the due list measures against. A program that sets only a cadence keeps the default template, and vice versa. When no program matches, the pair falls back to the tenant’s default rounding cadence setting (30 days out of the box). Surfaces show the program’s name, so a leader can see why they are being asked for a particular conversation on a particular clock.

Obligation covers direct reports; reach covers the subtree. A leader is anyone with direct reports. Their due list and coverage number count direct reports only — but they may manually round on anyone deeper in their reporting subtree (skip-level rounding), and doing so never inflates anyone’s overdue count.

Templates are copy-and-edit. Three starters install automatically — Staff Rounding, New Hire 30-60-90, and Senior Leader Rounding. Starters stay pristine: to customize, duplicate one and edit the copy (question types: Rating 1–5, Yes / No, Pick from a list, Free text, Recognition — pick a person, and Issue — logged to the Issue Ledger).

Pacing turns a long cadence into a short monthly list. A leader with 40 direct reports on a 120-day cadence should be rounding on about 10 a month, not 40. My Rounding shows a target for the current month — the sum, across the leader’s people, of how much of each person’s cadence falls inside the month — and an “up next” list of who to see, ordered by who has gone longest without a round. A leader can pin specific people to the front of the month’s list. Pacing changes only the view: whether someone is overdue is still computed from completed rounds, and coverage uses the same formula, so the two can never disagree.

Pillars are the tenant’s own strategy headings. Pillars (“People & Culture”, “Safety”, “Customer Experience”) are named by an admin, optionally from an industry preset. Template questions map to a pillar, so an issue raised from a question inherits it; a leader can override an individual issue’s pillar. Pillars group the ledger’s chips, the Themes tab’s sections, and the rows of the Stoplight Board.

Substantive rounds are what count. A completed round with every answer blank and no issue logged does not count toward coverage — that, plus subject visibility and same-day-cluster flagging, is how the app resists box-checking.

Themes roll up what leaders can already read. The Themes tab aggregates scale and yes/no answers by team, location, or job title across the viewer’s own reporting line. Every group with answers renders, whatever its size, because those rounds are attributed conversations the viewer can already open one by one.

Below that numeric breakdown, an admin can run AI themes: one clustering pass over the written answers of the last 90 days, producing labelled themes with a reach count and a suggested pillar. An admin can refile any theme under a different pillar, and that human call sticks — later extraction runs never move it. Unlike the numbers above them, AI themes are a tenant-wide aggregate, so the anonymity floor applies: a theme is listed only once it was drawn from at least the configured minimum group size (5 by default).

Team conversations let a leader hear from the whole team before rounding. When an admin switches them on, a leader can ask their direct reports three questions — what’s going well, what gets in the way, and what the leader should ask about on their rounds — in a short, anonymous conversation with an AI interviewer. Each person is invited in their Inbox, even if Leader Rounds was never rolled out to them, and can take part once. While it is open (1–14 days, 7 by default) the leader sees only how many have taken part. Once it closes — on its last day, or earlier with Close Now — the answers are put together into a readout: each topic shows what several people said, a topic fewer people than the anonymity floor spoke to is held back, and if fewer than the floor took part at all the leader sees only the count. A team smaller than the floor can’t be asked. Team conversations are off by default because every conversation uses AI credits.

Who Can Do What

Role	What they can do
Everyone	See the rounds about them and the issues they raised (“Rounds About Me”), with status and won’t-fix reasons; open any published Stoplight Board that covers them
Leaders (anyone with direct reports, or managers)	Capture and skip rounds on their subtree, manage the issue ledger for rounds they lead, see Team Rollup, Themes and the Rounds Log for their reporting line, compose and publish Stoplight Boards over their own org, and delegate their rounding upward
Admins / Leader Rounds app admins	Everything above across the whole tenant, plus Templates, Rounding Programs, Pillars, AI theme extraction, and Settings

Being the subject of a round grants read access only — a subject can never change the status of an issue they raised, even if they are themselves a leader.

Main Surfaces

• My Rounding — your due list (who’s past cadence, who’s never been rounded) and recent rounds, with Round Now and Skip actions. When team conversations are on, its “Hear from your team before your rounds” card starts one, shows the count while it is open, and shows the latest readout.
• Team conversations — every team conversation you opened, newest first, each with its readout (also on a phone).
• New Round — the capture form. Pick the person and template, answer as you talk, log issues with owners and due dates. You can also paraphrase the conversation into the AI assist bar and review the filled form. A “who deserves recognition” answer can be posted to Recognitions in one click afterward — the recognition goes through the normal moderation and approval gates.
• Issue Ledger — every issue from rounds you can see, filterable by stoplight color, paginated, with inline status updates.
• Team Rollup — coverage %, open reds, aging, and same-day-cluster flags per leader, with search, window filters (30/60/90 days), and CSV export.
• Themes — aggregated answers by team, location, or job title across your reporting line, plus the AI theme list an admin can extract from written answers.
• Stoplight Board — a red/yellow/green board with one row per pillar for a chosen leader and month, for huddle screens and slide decks. A leader composes it: the lights and issue counts are derived from that period’s ledger and cannot be hand-set, while the summary and next steps for each row are written by the leader (AI can draft them from the period’s issues, and the leader edits before saving). Publishing freezes the board — lights, counts and text are recorded as they stood — and adds a PDF export. Everyone the board covers can open a published board; drafts stay with their author and admins. A Leader Rounds admin can withdraw a published board, with a reason: it leaves everyone’s view and stays listed for admins as withdrawn, and the admin can start a corrected board for the same month from it. Once a new board for that month is published, an old link to the withdrawn one takes the people it covered to the new board.
• Rounds Log — every round you are allowed to see, not just the ones you led, filterable by leader and time window with the issue count and a substantive marker per round. This is how a director audits a coverage number: before it existed, a subordinate leader’s round was visible to their manager only if it happened to raise an issue.
• Rounds About Me — the subject tier: your rounds, your raised issues, and links back to the conversation each came from.
• Delegations — hand your rounding obligation to someone above you in the reporting line for a set window (parental leave, a secondment). It is a transfer: your people move onto the delegate’s due list and coverage number, and you stop being counted as due, for as long as the window is open. Only the obligation moves — your delegate could already see and round on these people — and a leader can hold one delegation at a time. Ending a delegation closes its window rather than deleting it, so past coverage still resolves correctly. Admins see who is covering for whom right now.
• Templates — the template library: three pristine starters plus your own copies, industry packs, the shared central library, and AI-drafted question sets.
• Rounding Programs — admin: which template and cadence applies to whom, as an ordered list of rules. Each rule shows how many people its audience actually matches, a rule that can never apply because one above it matches everyone says so, and a preview shows what a chosen leader’s team would resolve to.
• Pillars — admin: the strategic-pillar taxonomy, with industry presets offered on an empty list. Shows how many template questions map to each pillar. A pillar nothing references can be deleted; one with references archives instead, so history stays honest.
• Settings — admin: the fallback cadence, weekly reminders, red-issue escalation and its backstop, the monthly team digest, the anonymity floor, team conversations, the AI switches, and Load/Remove Sample Data.

Notifications & Automation

• Issue updates: when an issue you raised changes status, you get an in-app notification (deduplicated — one per status change), including the written reason on a won’t-fix.
• Weekly due reminders: once a week (Monday, 08:00 UTC), each leader whose direct reports are past cadence gets one in-app digest naming who’s due. Admins can turn this off in Settings; individuals can mute the Leader Rounds category in their notification preferences. A leader can also pause their own reminders from My Rounding.
• Red-issue escalation: an open issue still red a set number of days past its due date notifies the leader’s manager and re-nudges the owner, then repeats weekly while it stays red. Issues closed as won’t-fix are decisions and never escalate. Admins control both the switch and the grace window in Settings.
• Monthly team digest: on the 1st of each month, one aggregate “you said, we did” post per team on the News Feed — issues raised in rounds, issues resolved, what’s still open. It never names who raised what and only posts for teams at or above the anonymity floor. Off by default; the News Feed admin can also switch Leader Rounds contributions off separately.
• Team conversation invitations: when a leader opens a team conversation, each person on their team gets an invitation in their Inbox and as a push notification on the mobile app; it expires when the conversation closes. Everyone invited can switch these off under Leader Rounds · Team conversations in their notification preferences, including team members who don’t use Leader Rounds themselves. The leader is told in-app when the readout is ready.
• Home dashboard widget: leaders see rounds due and open reds on their dashboard (desktop and mobile), linking to exactly the lists counted.

The AI Assistant

The Leader Rounds assistant in Ask AI is read-only. It answers six kinds of questions: your due list, coverage rolled up your reporting line, ledger searches, the full history of one specific issue, your own rounds, and — when team conversations are on — what your team said in your own team conversations, never quoting anyone and never below the anonymity floor. It only ever reports what you can already see in the app, and private notes are not available to it at all. To capture a round, skip someone, or update an issue, it will point you into the app. Admins can disable the assistant in Settings.

Inside the product, AI drafts in three places, each with a human in the loop: describe-to-fill on the capture form, a drafted question set on a new template, and the Stoplight Board’s summary and next-steps text. In every one of them AI produces a draft that a person reviews and edits before anything is saved. Three more features send content to an AI model without drafting anything: voice dictation transcribes what a leader says on the capture form, theme extraction (run by an admin) clusters recurring themes from written answers, and — when team conversations are on — an AI interviewer talks with each team member and puts together the leader’s readout from what several of them said. Each has its own switch in Settings.

Settings

Setting	Default	What it does
Default rounding cadence	30 days	The fallback interval between rounds per direct report, used for any pair no Rounding Program covers. Range 1–365.
Weekly due reminders	On	The Monday in-app digest to leaders with due rounds
Escalate aging red issues to the leader’s manager	On	An open issue still red past the grace window below notifies the leader’s manager and re-nudges the owner, weekly while it stays red. Won’t-fix issues never escalate.
Days an issue stays red before escalating	7 days	The grace window, counted from the issue’s due date, so a leader always gets this long before anyone above them hears about it. Range 1–90.
Post a monthly “you said, we did” digest to each team’s feed	Off	One aggregate News Feed post per team on the 1st of the month. Never names who raised what; only posts for teams at or above the anonymity floor.
Minimum group size (anonymity floor)	5 people	Applies in five places: the monthly team digest posts to a team only once this many distinct people had a substantive round; a newsletter block scoped to a single team shows numbers only at or above this size; on the Themes tab an AI theme is listed only when it was drawn from at least this many people; and on a Stoplight Board a pillar row built from fewer than this many distinct subjects keeps its light but hides its counts, in the published PDF too; and in a team conversation’s readout a topic is shown only once this many people spoke to it. It does not apply to the numeric answer breakdown on the Themes tab — those rounds are attributed and the tab is visible only to leaders of the people in it. Raising this hides existing themes and board rows; lowering it never lowers the floor a team was promised when its conversation opened. The lowest value is your company’s smallest group (set in Privacy), up to 25.
Escalation backstop	None	A recipient group that hears about a stuck red issue when its leader has nobody above them. Left blank, those escalate to nobody.
Leader Rounds AI assistant	On	The read-only Ask AI assistant
Team conversations before rounds	Off	Lets leaders open anonymous AI conversations with their team. Every conversation uses AI credits.
AI theme extraction	On	The Extract Themes run on the Themes tab
AI drafting	On	Describe-to-fill on the capture form, the Stoplight Board’s drafted text, and Generate with AI on Templates
Voice dictation	On	The Speak button on the capture form, which sends the recording for AI transcription

Settings also carries Load/Remove Sample Data, which seeds a full demo state (rounds, every stoplight color, a same-day cluster) and removes it exactly.

Scope

Leader Rounds is staff rounding only: leaders rounding on their people. Patient rounding, EHR/census integration, and HCAHPS are deliberately out of scope. Historical import and offline capture are deferred by design.