📥 My Work

My Work answers one question: what actually needs me today? Everything waiting on you across the whole platform is gathered into a single list, so you never have to check each module in turn and hope you have not missed one.

The attention counter

The badge in the header reads immediate / total — how many items need action now, out of everything not yet finished. When it clears, you are genuinely up to date. The sidebar carries the same count.

What appears here

Awaiting your review, your submissions due, returned to you, your open action items, and meetings needing attention. Sections only appear when they have something in them, so an empty queue really does look empty rather than showing five headings with nothing under them. Every row links straight to the screen where you can deal with it, and the item drops off the list once it is done.

Long queues

Each section shows the most relevant handful of items while its summary tile counts the whole queue. Where the two differ you see a Showing 5 of 23 note, so a long list is never silently truncated into looking complete. The action-item and meeting sections link out to their full screens.

My Action Items

The action-item side has its own screen at My Action Items, with four ways to look at the same list — a compact list, a sortable table, a due-date calendar, and a board grouped by status. It is covered in full under Meeting Notes AI.

📈 Ups & Downs

The Ups & Downs module is a structured cycle review workflow. Each project has a recurring cycle (weekly, bi-weekly, or monthly) during which Project Leads submit account-level highlights (Ups) and risks (Downs).

Approval chain

Submissions travel up a fixed 6-level chain: PL → PM → Delivery SPOC → Sales SPOC → POD Owner → Cluster Head. Each reviewer can approve and forward, or return with a comment.

The advantage over a status report is that you write it once. A risk a team raises on Monday becomes visible to cluster leadership without anyone building a deck for it, and if something needs rework a reviewer returns it with a comment attached rather than starting an email thread.

Your role in this workflow

• 📈 Project Lead — you start each cycle: compose your Ups & Downs and Submit to Project Manager →, or 💾 Save as Draft to continue later. You cannot edit after submitting — ask your PM to return the cycle if you need a change.
• ✅ Project Manager — review your PLs’ submissions, optionally add your own items (labelled “Added by Project Manager”), then Approve (forwards to the Delivery SPOC) or Return (back to the PL with your comment).
• 🚚 Delivery SPOC — second-level review. Approve forwards to the Sales SPOC; to send it back, choose the level (see below).
• 🤝 Sales SPOC — review after the Delivery SPOC, adding client / commercial context. Approve forwards to the POD Owner.
• 📦 POD Owner — review cycles across every project in your POD; Approve forwards to the Cluster Head.
• 🏛️ Cluster Head — final approval. Approving marks the cycle Cluster Head Approved (closed). You also see the full Review Chain Summary.
• 👁️ PMO Observer — read-only oversight across all projects and cycles, with no approval actions.

Returning a cycle. Above the PM, you get one Return to … button for each level below you, so you choose where the cycle goes. You can return it while it is still with a lower level, not only once it reaches you; a button is greyed out when the cycle is already at that level or has not reached it, and the page says why. Returning to a reviewer puts the cycle in their queue. Returning to the Project Lead asks them to resubmit, and their resubmit goes straight back to the level the cycle was waiting on when you returned it, so no review is skipped and the approvals below it are kept.

While a Project Lead is still to submit, the PM’s Project Lead Submissions panel says how long is left, such as Closes in 2d 4h, or that the deadline has passed. The Dashboard shows an Ups & Downs Awaiting You banner when cycles are waiting at your step; it opens My Work. The Cluster Head screen has a Portfolio — final approvals card listing every project in your scope with its latest cycle, the ones awaiting your final approval first. The POD Owner screen has the same card, titled POD overview, and each cycle on it says how long it has been at its current step.

On the Delivery SPOC, Sales SPOC, POD Owner and PMO screens, 📄 Export PDF downloads the cycle you are reviewing as a PDF, without going to Cycle History.

Adding items. Type each Up or Down in the text area at the top of its card and click Add Up or Add Down. Repeat for each item; the formatting toolbar gives bold, italic and bullet lists.

On the Projects screen you can narrow the list by POD and by cycle type.

Leaving a comment when you review. The comment box is there whether you approve or return. Under it, Add a reason adds a standard return reason in one click, and 🤖 Draft the return with AiDE suggests which items to revise and writes a draft note into the box, building on the reasons you added. It is only a draft: nothing is sent until you edit it and press Return. It uses your own AiDE Fabric Token.

How a cycle is labelled depends on how often your project runs one — its frequency is set per project. Every label is followed by the dates the cycle covers.

Frequency	Label looks like	What it names
Weekly or bi-weekly	W23-2026-W23 (1 Jun – 7 Jun)	The ISO week the cycle starts in
Monthly	SEP (M09)-2026 (1 Sept – 30 Sept)	The calendar month — the whole month is one cycle

Older cycles can still show an earlier style with a month name in front, such as JUN-2026-W23. It is the same kind of cycle, labelled before the current format was introduced — the dates in brackets are what to go by.

🛰️ AiDE Risk Radar — Cluster Heads

On the Cluster Head review screen there is an AiDE Risk Radar card. It reads the Down items from recent cycles and the meeting action items that are past their due date and still not done across your projects, and ranks them into one short cross-project list — so you can see where attention is needed without opening each project in turn.

Each row names the project, a severity — red, amber or green — a short description of the concern, and a trend arrow showing whether it is getting worse, improving, or holding steady. The card also tells you how many projects were scanned to produce it.

It covers only the projects where you are the Cluster Head; there is no project picker. It does not run when you open the screen: each run is a paid AiDE call, so it waits until you press the button on the card. If you have run one before, the card shows that earlier result with the time it was generated at — a snapshot, so generate again for today’s picture. It needs your AiDE token to be connected (Profile).

Getting it by email. The same radar can arrive as a weekly leadership digest to every active Cluster Head, instead of being something you remember to open. It is off by default and an administrator turns it on in Admin Portal → Scheduler; each digest spends one AiDE call.

Next-cycle pre-planning

A Next Cycle tab lets a Project Lead start drafting the following cycle before the current one closes, so nothing is lost between windows. It stays a private draft until the cycle is created and announced — 14 days before its end date, when the kickoff email goes out. From then it can be submitted, and it is due at the end of its last day.

Cycle History

Cycle History browses every closed cycle for any project you have access to, and opening one gives a detail view with the full item list and the activity trail — which is what makes the archive worth keeping.

Once a project has two or more cycles, a chart above the list shows its Ups and Downs per cycle, oldest on the left, so you can see whether delivery is improving or slipping. It covers the cycles listed and says when those are only the most recent ones.

Use the project and status filters to narrow the list. A cycle a reviewer sent back has its own Returned status option, and In Progress does not include it, so the filter agrees with the Status column.

A closed cycle is read-only. A cycle still in progress that is waiting on a level you may review has a ⚡ Review Action card, so you can act on an older cycle as well as the latest one: Approve & Forward or Return to …, the same as on your review screen. The confirmation names the reviewer the cycle actually reached, which can be a level further on when nobody is active at the one in between.

In the Approval Chain, each level’s cycle time has a small bar sized against the slowest level, amber where it held things up.

Editing your project

On the Projects list, ✏️ Edit project on a project card opens the project for its Project Lead and Project Manager (and for admins). What you can change depends on your role, and the dialog shows only those fields:

Role	Can change
Project Lead	Name, code, cycle type and status (with an optional reason, kept with the change), the note for the next kickoff email, the Contributors and the Project Leads.
Project Manager and the senior roles	The Contributors only. The rest is shown read-only. A Delivery SPOC, Sales SPOC, POD Owner or Cluster Head has Edit project on every project in their list, and can also open the same dialog with Manage contributors under People & Organisation.
PMO Associate and System Admin	Everything, as on Admin → Projects.

The Note for the next cycle kickoff email goes out once, in the next kickoff email to the project’s Project Leads and Managers, and then clears. Project Managers, the Delivery SPOC, Sales SPOC, POD Owner and Cluster Head are set by the PMO, because they decide who reviews your work.

Temporary access to a project

You can let somebody outside the team use one project for a while. They can see its meetings and log their own weekly updates against it; they cannot review anyone, and they do not become a member of the project. The Temporary access panel is in the project dialog: ✏️ Edit project on the Projects list for the Project Lead, the Project Manager and admins, or Manage contributors under People & Organisation for the senior roles.

Give access: search anyone by name or email, pick the person, choose the date the access ends (at most 90 days ahead) and, if you like, say why. Then press Give temporary access. The access lasts to the end of the day you chose.

Invite someone with no account: type their work email. If it matches nobody, ✉️ Invite … — no account yet appears; choose the end date first. They get an email saying who invited them and to which project, and the access is waiting when they accept. Until then they cannot sign in or see anything. If they do not accept in time, the invitation is sent once more with a fresh link, and only once. Work addresses only: anyone outside the company goes through the PMO.

Extend or withdraw: each grant shows ✅ Active, ⌛ Expired or 🚫 Revoked. On live access, Extend moves the end date later; anyone who leads or manages the project can do it, not only the person who granted it. The new date can be at most 90 days from the day you extend. Withdraw ends the access now; extending never shortens it.

🎯 Priorities & Achievements

The P&A module captures weekly priorities and achievements for each project lead (and contributors), an AiDE adoption entry, and travels through a bilateral review chain.

Submitting your weekly P&A

Open Priorities & Achievements from the sidebar — the current active week is pre-selected, and if no week is open a message tells you when the next one does. Click + Add Priority, type it in the rich text editor, and choose a category. Achievements are added the same way with + Add Achievement, for the previous week. A priority worded exactly like one of yours from last week is marked ↩ Carried over from last week for you and your reviewer — a note, not a judgement.

Just above the Achievements section, last week’s priorities and achievements are shown for reference, so you can see what you committed to. If you missed last week, your most recent submitted week is shown instead, with its dates and a note that last week was not submitted.

If a reviewer returns one item. A reviewer can approve or return a single item as well as the whole week. A returned item shows ↩ Returned by the reviewer and why, under the item. You can change that item with ✏️ Revise while the rest of your week stays submitted. When you save, the reviewer is told and the line reads ✎ Changed since they looked at it. Until you revise a returned item, your week waits on you: it cannot be approved, acknowledged or marked as noted, and it is not approved automatically. Reviewers can still decide your other items (v849, #5290). If the whole week is returned instead, change what you need and submit again.

Reviewing your team’s P&A

As a reviewer (PM, Delivery SPOC, Sales SPOC, POD Owner or Cluster Head) the same screen shows a team-level view with one card per submitter. Click a name to expand their submission; the card carries the submission time, the categories, and your approve / return controls. Where you are the first reviewer for an item, its row also has ✓ Approve item and ↩ Return item (a return asks for a reason, and the person is told in their notifications, not by email). When they change that item with ✏️ Revise, you are told, its badge reads ✎ Revised, look again, and you can decide it again; if someone had already approved it, their approval was of the old text, so it waits again too and they are told as well. This is separate from approving or returning the whole week. While an item you or another reviewer returned waits on the person, the week cannot be approved: Approve is hidden and a note says why. You can still decide the other items (v849, #5290). A Delivery SPOC, Sales SPOC, POD Owner or Cluster Head on the work can ⚡ Override-approve a stalled week, with a reason that goes in the activity trail.

On the same screen, 📊 Your Team’s P&A Status lists everyone expected to submit, under each project you oversee, for this week and the two before, with a line saying how many have not submitted this week. Someone who has not started is listed as not submitted. Click a status to open that week.

To approve several at once, press ☑ Select on the review queue, tick the submissions awaiting you, then Approve selected. The confirmation names everyone and warns how many you have not opened in this session; each is still your commitment to that person. Tick Compare with previous week to see, beside each member, what they recorded the week before and how their counts changed. The Words per entry card and badges show how detailed entries are (under five words is marked Brief) — a pointer, not a grade.

Cluster Head weekly digest

Cluster Heads see P&A Digest in the sidebar instead of the standard review screen — every submission from every team member in one consolidated view, grouped by project and role. Acknowledge records that you have read the week. A member whose week was returned shows ↩ Awaiting Resubmission instead, and one with a single item returned shows ↩ Awaiting Revision; you can acknowledge them once they have done it (v849, #5290).

P&A History

P&A History reviews any past week: pick a week from the selector and browse every submission, its categories and its approval status. PMO Associates can view submissions from all team members.

Approval chain

The P&A chain follows the project or POD each item is tagged to — not a separate org chart. The first reviewer is the nearest person above you on that project, and the chain steps up the role ladder to Cluster Head level. Empty rungs are skipped automatically — if there is no PM on the project, the chain goes straight to the Delivery SPOC.

If you are on no project, you can reach no POD either, so there is nothing of your own to tag an item to. When an admin has set a default POD, that POD is offered to you and your items go to its reviewers. Until one is set, you can still submit without naming a project or POD. Anyone with projects of their own sees only their own projects and PODs.

Your P&A is due by the end of Monday in your own timezone. Reminders go out Monday to Friday until you submit, and from Tuesday they say the week is overdue. Your on-time record on the Profile page shows how often you have made the deadline.

If your first reviewer does not review within 36 hours

When nobody at the first review step on your submission has approved or returned it 36 hours after you submitted, the system approves it for them, whoever the first reviewer is: your Project Lead if you are a Project Contributor, your Project Manager if you are a Project Lead, or a more senior reviewer when there is no one below them (v865, #5402). It then shows as 🤖 Auto-approved: accepted, like an approval, but given by no person. You get an email saying so, and each first-step reviewer who did not review it gets one too. While it waits, your P&A page shows when that will happen, beside the Submit button — and shows nothing when your submission is not one the system would ever approve. The Project Manager sees the same time on their review page, under the feedback box, until they approve or return it.

🗓️ Project Timeline

The Project Timeline is a shared, per-project running log where any team member can post events (milestones, blockers, announcements), note revisions, and capture items directly into Ups & Downs or P&A drafts.

Who can see what

Event visibility is controlled by the Visible to everyone box: 🔒 Chain (default — unchecked), the author and everyone at or above them in the reporting chain who is on the project, or 🌐 Everyone (checked), all project participants. PMO Associates always see every event as read-only observers regardless of the setting. The 👁 N can see badge on each event shows how many people that is; hover it for the full roster, sorted by seniority.

Posting an event

Pick the project from the Project dropdown at the top of the content area — only projects you have access to are listed, and the feed reloads when you switch. Compose in the Add an event box, choose a type, optionally back-date it, optionally pick the Related meeting it is about (the project’s meetings and its POD’s, which are marked) (the card then opens that meeting, and the meeting lists the post), tick Visible to everyone if it should not follow the chain, and attach files before posting. The capture chips under the composer copy the event into an Ups & Downs or P&A draft as well; then press Post to Timeline.

Editing an event

If you are the author, or at the same role level or above in the reporting chain, ✎ Edit / add info opens an inline form under the event body, where you can also change or remove its Related meeting. Previous versions are kept, and if the event has drafted captures, editing the body updates the linked Ups & Downs or P&A draft too.

Deleting an event and its notes

Delete soft-deletes the event and takes its notes, revisions and attachments with it — but not the linked draft items in Ups & Downs or P&A, which you must remove separately. A note can be deleted only by the person who wrote it; the event author cannot delete someone else’s note.

Attachments

Every event card carries a file section at the bottom — screenshots, documents, spreadsheets, anything that is supporting evidence — and files can be added to an event that is already posted. Deleting an attachment is permanent, and only the person who uploaded it or the event’s author can delete it; a manager who can edit the event can add files but not remove yours.

Timeline views

The toggle above the feed switches between 🗓️ Timeline — the event stream grouped by day, newest first, with back-dated events sorted by their back-date — and 📋 Activity log, a flat chronological list of every action: who posted, edited or deleted what, and when.

From and To dates narrow the feed to a date range. Escalations, blockers and risks, the entries Risk Radar reads, carry a Flagged badge, and the type dropdown has a Flagged option for all three.

🤝 Meeting Notes AI

Meeting Notes AI turns what a meeting produced into structured minutes. You upload a transcript document, paste the text, or upload a recording (video or audio) — a recording is transcribed for you first — and AiDE generates the minutes of the meeting, a topic mindmap, and a list of action items with owners and due dates. You then track those actions to closure and share the outcome with participants.

Who can use it

Any ValueLabs employee with a POD Governance account can create meetings. A meeting is visible to its owner, direct participants, and — when you toggle Visible to everyone — every member of the project team.

Reading the meetings list

Four tiles summarise every meeting you can see: Meetings visible to me, Open action items, Overdue action items and your Action closure rate. Each row carries badges:

Badge	Meaning
Mode	💻 Teams, 🎦 Zoom, 🏢 In-person or 🔀 Hybrid
Status	📥 Awaiting transcript, 🤖 AiDE processing… or ✅ Processed
Audience	🌐 Everyone, or 🔒 N people (participants and the owner only)
Overdue	⚠ N overdue when open action items are past their due date
Closure	A progress bar and “N/M closed”

Connecting your AiDE Fabric Token

AiDE processing requires a personal AiDE Fabric Token (aide_fabric_token_…) obtained from aide.imagine.tech ↗. Enter it once via the Connect AiDE banner on any meeting — it is stored encrypted against your account and only a masked preview is shown. You can also replace or remove it from My Profile → 🔑 AiDE Fabric Token, which connects it for the portal’s other AiDE features too.

If AiDE refuses your token, most likely because it expired, it no longer shows as connected, so you can replace it before a feature fails.

Create a meeting and add the sources

Click New meeting and give it a title, mode, date, time and duration, and optionally participants. On the meeting’s page add what it produced — a transcript document (.docx, .txt or .vtt, up to 10 MB), pasted text, a recording (video or audio), or a mix — then click Process with AiDE. For a recording, choose Generate transcript and it is turned into speaker-attributed text first. Uploading a recording is for Project Lead and above; anyone who can process the meeting can add a document or paste text.

A recording can be a video or an audio file: .mp4, .mkv, .webm, .mov or .m4a, up to 50 MB. That includes a file from your phone’s voice recorder app, such as Voice Memos on an iPhone.

You can also record the meeting on the meeting page itself. Under 🎙️ Record the conversation, press Start recording; your browser asks to use the microphone the first time. You can Pause and Resume, and Stop & attach adds the recording to the meeting, ready for Generate transcript. A bar shows how much of the 50 MB you have used. At the limit the recording stops and is attached on its own; start a new one to carry on, as a meeting can have several. If a recording was cut off before it could be attached, the page offers it next time: Attach it or Discard. The recorder only shows in a browser that can record.

Before you generate a transcript, choose the Spoken language: Auto-detect, or one of the listed languages, including Hindi, Telugu, Tamil, Kannada, Malayalam, Marathi, Bengali, Gujarati, Punjabi and Urdu. Naming it gives better transcripts for short calls, strong accents and meetings in Indian languages. The choice is saved on the meeting, and the next meeting in a series keeps it.

Recurring meetings, occurrences & project link

A meeting becomes the first of a series when you turn on 🔁 Recurring series while creating it and choose how it Repeats: weekly, bi-weekly, monthly or ad hoc. Then use ＋ Add occurrence for the next one. Each occurrence copies the participants but has its own transcript and outcome; the series lists them all. Link a meeting to a project you can open, and sharing its outcome drops a card on that project’s Timeline linking back (a restricted meeting posts a title-less record).

On Meeting Notes AI, a series’ occurrences are grouped by month and then by week, newest first, and each group can be folded. Each group shows how many meetings it holds and how many of their action items are closed, and each occurrence has its own progress bar. The newest month and week are open when you expand the series. A year heading appears only when the series runs across more than one year.

Series owners and co-owners can take someone off a series in one step: Remove from upcoming occurrences on any occurrence. The person comes off every upcoming meeting and stays off new ones; meetings that already happened keep them. Anyone with an open action item on an upcoming meeting has to hand it over first, and adding them back undoes the removal.

A series owner or co-owner can use ⏸ Pause series… to stop new meetings until a date; the first meeting after it falls on the series’ usual day, and ▶ Resume now ends the pause early. ⏹ End series… picks the date of the last meeting: no more are added after it, past meetings and their minutes are kept, and calendar invites stop repeating on that day. ▶ Continue series takes the end date away. Owners also see who has been involved across the series’ past meetings, least involved first; it counts who was listed on each meeting, not who attended. While the series’ project or POD is inactive, no new meetings are added. Once it is active again, the next meeting falls on the series’ usual day.

On the Action Items tab your own items are tagged You, and My items only shows just yours. Ticking an item done asks for a justification, a short note on what was done, which is recorded on the item; re-opening one asks for nothing. Access follows the participant list, and Delivery SPOCs, Sales SPOCs, POD Owners and Cluster Heads can also open, edit and process the meetings on the projects and PODs where they hold that seat, the same as a co-owner.

Meeting lifecycle

Stage	What happens
Draft	Created but no transcript uploaded yet. You can add participants and set the mode.
Processing	Transcript staged; AiDE is generating the summary, mindmap and action items.
Ready	AI output available. Review and edit before sharing.
Shared / Finalised	Report locked and distributed to participants (or everyone). Action items enter the reminder cycle.

Action-item reminders

Assignees receive an email three days before an action item is due, again on the due date, and every two days while overdue (the meeting owner is copied on overdue escalations). Reminders can be paused per meeting via the 🔔 toggle on the meeting detail page. If nobody said when an action is due, it is due the next day, the same as one you add yourself; change the date on the item. The tab shows how many are closed out of the total. When AiDE found no tasks with an owner or a due date, it says so: that is a finished extraction, not a failure, and you can still add one by hand.

My Action Items

Closing several at once. In the List or Table view, press ☑ Select, tick the open items you have finished (or Select all), then Mark done. You write the justification once and it is recorded on each item, as if you had closed them one by one; anything that cannot be closed is listed with the reason.

🎤 Voice add, next to New action item, lets you speak an item instead of typing it. The first time, your browser asks for the microphone — click Allow; the button changes to 🎙️ Listening…. Say what needs doing in one go; the New action item form opens with your words in it, ready for a due date.

🔁 Recurring tasks

A task you raise yourself can repeat. In New action item, give it a due date, then pick a Repeat: Daily, Weekdays, Weekly, Monthly, Yearly, or Custom… (for example every 2 weeks on Mon and Thu), with an optional end. Each date becomes its own task, created 7 days before it is due on your own calendar day, and shows 🔁 with its rule in plain words.

You want to	What to do
Change one task only	Open it, edit, and choose This task only.
Change this task and the ones after it	Choose This and future tasks. Every open future task is updated to match, including ones somebody has already worked on.
Stop it	Open any of its tasks and press ⏹ Stop repeating. Tasks already created are kept.
Stop one somebody else set up for you	Press ⏹ Stop repeating on the task in My Action Items. You cannot change its rule, but you can always stop it, and the person who set it up is told.

If the task’s project or POD is made inactive, no new tasks are made and the open ones are not reminded. Once it is active again the next one appears as usual; the dates that fell while it was inactive are skipped, not all created at once.

Focus timer

The panel at the top of My Action Items is a pomodoro timer that records the time you spend on a specific action item. It is built around one rule: focus time is always recorded against a task, so the timer will not start until you have chosen one.

Key	What it does
Space	Start the timer, or pause and resume a running one.
R	Reset the current stretch.
S	Skip to the next phase.
Z	Zen mode — the dial on its own, without the rest of the screen.
Escape	Leave zen mode.

When you reach your daily goal the goal card says so, once. Turn on Daily focus goal reached in My Profile’s notification settings to get a bell notification too (and a push where push is on) — never an email, at most once a day. Your history is on My Profile in the 🍅 Focus Activity card: the last 91 days, your all-time totals, and your streak — days in a row with at least one finished pomodoro, in your own timezone.

🩺 Project Health

Project Health writes the account summary you would otherwise assemble by hand — reading what a project has already reported over a period, so nobody is asked to report anything twice.

One screen, three scopes

The buttons at the top of Health choose how wide the summary reaches:

Scope	Covers	You see it when
Project	One project, for a week or a month	Always
POD	Every project in a POD, in one read	At least one POD has a project you can read
My portfolio	Every project you are responsible for, across every POD	You are responsible for at least one project — the button shows how many

Where the content comes from

Three governed sources you already have access to: Project Timeline events, Meeting Notes AI outcomes, and Priorities & Achievements. Nothing is invented from outside the record, so every point traces back to something you can open and check.

Reading the output

An executive summary with highlights and risks pulled out separately, plus a health score and grade with a score delta against earlier periods (so you read direction, not a bare number) — and the rationale behind that score. Below it, Tabbed detail lists the achievements, meetings, risks and highlights with their underlying items. The score is produced by AiDE as part of its analysis, not a formula applied afterwards, which is what lets the rationale genuinely explain it.

Sharing and scheduling

Use 📋 Copy to paste it into a message, ✉️ Email to send it, or 📄 PDF to export it for a pack. ✉️ Email then tells you the summary has been emailed to you or, when it was not sent, why (for example you have turned that kind of email off). The scheduled delivery card turns on automatic weekly or monthly generation for that project, so it arrives without you having to remember. Treat a summary as a strong first draft, not a finished document.

Apply to all my projects sets the same delivery for every project you can pick, instead of one at a time.

Generating a summary

Only projects you can already see are listed — if the list is empty you are not yet mapped to a project, so ask your administrator. Weekly gives you the detail for a review conversation; monthly steps back for a trend view, which suits a steering committee better than a stand-up. You can also pick a past period.

Access and prerequisites

You need a project assignment and a connected AiDE Fabric Token — the same token that powers Meeting Notes AI processing. If the screen reports a token problem, connect one from My Profile. Summaries also surface as suggestions inside the Ups & Downs and P&A screens.

📊 Reports & Analytics

The Reports screen is available to PMO Associates, Delivery SPOCs, POD Owners, and Cluster Heads. It shows cycle statistics, Priorities & Achievements trends, AiDE adoption, return rates, and notifications delivery metrics. What it covers depends on your role. A PMO Associate sees every project; a Delivery SPOC the projects they hold; a POD Owner or Cluster Head every project in their POD — the same list the project picker offers, so if your numbers are smaller than a colleague’s, that is why.

Filtering — date range, POD & project

A toolbar at the top re-scopes the whole page: 📅 Date range (All time, 7d, 30d, 90d or a custom from/to — the headline “Total Cycles” switches from its year-to-date default to your window), plus 🏢 POD and project narrowing.

Trend deltas & drill-downs

There are eight headline cards: Total Cycles (YTD), Approved Cycles, In-Progress, Priorities & Achievements Completion (the share of submitted P&A reviews that have been approved, organisation-wide even under a filter), Open Action Items, Return Rate, Avg Cycle Time and AiDE-assisted P&A.

Each headline card carries a short trend delta — “▲ 3 this week” under Total Cycles, “▲ 2 this month” under Approved. Those are fixed-period indicators and do not follow the date filter. Lower down, By POD & project breaks completion and return counts down per project.

The charts

• Ups & Downs trend — a pair of bars per cycle, Ups (blue) and Downs (red) side by side, with the net figure in the table; toggle between the chart and the underlying table with the Table / Chart switch.
• Return rate — how often cycles were returned at each stage (PM → PL, Delivery SPOC → PM …). A high rate at one stage points at a quality or communication problem there; hover a bar for the most common reason.
• P&A completion — the weekly submission rate as a line. The SLA card below it gives average, minimum and maximum approval times and the share of cycles approved inside the 3-day target.

More report panels

• Everyone with Reports: AiDE insights (comments on the headline numbers and a plain-words question box; both call AiDE, nothing is applied on its own), Cycle Time by Stage (median and P90 per approval stage), Achievement Quality (weekly average AiDE score, from saved scores), AiDE Priority Adoption, the Meeting Notes AI report, Action Item Ageing (with on-time finishing this year for the ten assignees with most open), Project Timeline Activity and Evidence Attachments.
• PMO / Admin only, because they count the whole organisation: Focus Timer Adoption, User Adoption & Access, Email Deliverability, Notifications & Digest Delivery and MCP Plugin Usage.

CSV exports & scheduled reports

The ⬇️ Data Exports card has one button per CSV file: All Cycles, Ups & Downs Trend, Return Rate by Stage, Cycle Time (SLA), Governance SLAs & Participation, Meetings & Action Items, Attachments Register and Action Item Ageing for everyone with Reports, and for PMO and System Admins also Priorities & Achievements History, Audit Log, Digest Delivery Log, User Adoption & Access, Email Deliverability and MCP Usage Log. Project data covers only the projects you have access to. Every file is formula-safe — any cell that would trigger a spreadsheet formula is prefixed automatically, so opening one in Excel cannot execute anything. PMO Associates and System Admins can schedule reports from Admin → Reports & Exports, emailed weekly or monthly as a CSV attachment (there is no daily cadence and no PDF attachment).

Cycle time, governance SLA, user activity, the Ups & Downs trend and the digest runs can be scheduled too, and each covers the schedule’s own period, such as the last 7 days for a weekly one. For PMO and admins, the User Activity card shows sign-ins, submissions and approvals per person over the date range you pick, so you can see who has gone quiet lately.

🔔 Notifications

Notifications keep you informed about everything that needs your attention — cycles awaiting your review, Priorities & Achievements returned for revision, weekly digests, approaching deadlines, and bilateral approvals. They appear in two places: the 🔔 bell in the top bar (a quick dropdown of the latest items) and the full Notifications screen (the complete, filterable history).

The notification bell

The bell in the top bar shows an unread count badge. Click it for a dropdown of recent notifications; click any item to jump straight to the relevant cycle, Priorities & Achievements week, or report. A 🔕 icon means notifications are currently snoozed.

The Notifications screen

Open Notifications from the sidebar for the full, day-bucketed feed. Each row shows the category, title, message, and time. Unread rows are tinted light blue.

Action	What it does
Click a row	Marks it read and navigates to the related item
Mark all read	Marks every unread item read — with a 5-second Undo if you change your mind
Unread only	Toggles the list to show only items you have not yet read
Category filter	Narrows the list to one category (Ups & Downs, Priorities, Meeting Notes AI, Digest, Timeline, Account) — while one is active, Mark all read becomes Mark <category> read and only clears what is on screen
Clear filters	Resets the category filter and the unread-only toggle
✕ on a row	Dismisses that notification, so it leaves your list and your unread count
Clear read	Dismisses every notification you have already read

Notification categories

Category	When you get it
📊 Ups & Downs	A cycle is waiting for you, moved on, was returned to you, or completed its chain, and the reminders about it
🎯 Priorities	A P&A week needs your review, yours was approved, returned or auto-approved, or a single item was approved or returned
🎤 Meeting Notes AI	A meeting’s outcome is ready or finalised, an action item was assigned to you, is due or overdue, you reached your daily focus goal, or someone stopped a repeating task you set up
📧 Digest	Your weekly digest, your daily briefing, the leadership digests, the stalled-cycles alert, and your weekly focus digest
📅 Timeline	An event was posted or deleted on a project timeline you are part of
👤 Account	An invitation, your project access was given, removed or changed (including temporary access), your role changed, an access request was made or answered, or a leaver’s work was handed to you

The same six chips filter the list, and the URL carries the filter (?category=Digest&unread=1), so a filtered view can be bookmarked or shared.

---

Preferences & snoozing

Control which notifications you receive — and how — from the Notification Preferences card on your Profile. You can mute individual project timelines and snooze all notifications for a set period (the bell shows 🔕 while snoozed). Email notifications mirror the in-app feed and respect the same preferences, including the weekly digest frequency (Weekly, Daily, or Off). The opt-in Weekly focus digest arrives on Friday afternoons. Every notification email also carries links to stop that kind of email or to manage your settings without signing in; the page they open changes nothing until you click, and sign-in and account emails always arrive.

Desktop & mobile notifications (push)

Install POD Governance as an app and turn on push to get desktop and mobile alerts, and your count on the app icon, even when the app is closed. The bell still shows everything; push is optional and set per device.

1. Install the app

Push works on the installed app, not a browser tab: Chrome / Edge on a computer (install icon in the address bar), Android Chrome (Add to Home screen), or iPhone / iPad Safari on iOS 16.4 or later (Share → Add to Home Screen). See Installing POD Governance as an app under Getting Started.

2. Turn on push

Open Profile → 🔔 Push Notifications (this device) and click Enable on this device; your browser asks for permission only then. Send me a test push checks it works, and per-category switches choose what pushes (separate from your email switches). The Controlling your notification preferences video under Profile & Tokens shows the same Push Notifications card.

The app-icon badge

A number on the app icon shows your My Work “immediate” count (overdue action items, returned submissions and long-waiting reviews) when it is above zero, otherwise your unread count, and clears once nothing is overdue and everything is read.

Steps to make the count appear

It needs the app installed and push allowed. On iOS specifically, no badge shows until you grant permission. Desktop (Chrome/Edge/installed Safari) and iOS/iPadOS 16.4+ show the number; Android shows only a dot, which is the platform’s limit — the bell and My Work always show the exact count.

Browser support

Browser / platform	Push	Badge
Chrome / Edge — Windows, macOS	Yes	Yes — numeric
Chrome — Android	Yes	Dot only (Android limitation)
Safari — macOS 13+ (installed)	Yes	Yes — numeric
Safari — iOS/iPadOS 16.4+ (Home Screen, permission granted)	Yes	Yes — numeric
Firefox — desktop	Yes	No (push only)
Firefox — Android	Limited — best-effort	No
Any browser tab (not installed)	Limited / no	No

📋 Audit Log

Every state-changing action in the portal is recorded in the append-only audit log. Admins see the full app-wide log; other roles see the log filtered to their projects and submissions.

What is recorded

Cycle and P&A state changes, item edits, user and project administration, token issue and revocation, and email sends. Three things are deliberately not audited, to keep the log readable: notification read-receipts, the system health endpoint, and the dev-only scheduler reset.

Reading an entry

An entry names the actor and the role they were acting as, the action key (cycle.approved, cycle.item_edited …), the entity it touched, a plain-language description, and a UTC timestamp. Where a field actually changed, the before/after diff is carried with it; a pure state change says so rather than showing an empty diff.

Filtering the admin audit log

The filter bar on Admin → Audit Log has a search box for the person, the action or the project, and filters by entity (cycle, priorities, user, project …), role (the role the actor was using), event type, and From and To dates. The filtered result exports to CSV.

🔌 MCP Integration

POD Governance ships an MCP (Model Context Protocol) server that lets AI agents (Claude, GitHub Copilot, etc.) query and update the portal via natural language.

Generating a Personal Access Token

You need a PAT to authenticate your AI agent with the MCP server. Option A — Self-service (recommended for individuals). Go to My Profile → MCP Access Tokens, click Generate Token, copy the pgmcp_… value (shown only once), and use it as MCP_API_TOKEN or paste it into a remote-server form.

Option B — Admin-issued (service accounts & oversight). Admin → MCP → Tokens → Generate, on behalf of a service identity. Keep the value in a secret vault and never commit it.

An admin can also limit a service account's token to named MCP tools, when creating it or later with Tools… on the token's row. The token is then refused for any other tool, so an automation that only reads cycles cannot reach meetings or approvals. This keeps an automation to what it needs; a leaked token is still contained by the account's role, the token's read or write scope, and revoking it.

MCP token vs AiDE Fabric Token

They point in opposite directions and are never interchangeable. The pgmcp_ MCP token lets an assistant call in to POD Governance as you; the aide_fabric_token_ Fabric Token lets the portal’s own AiDE-powered features (Meeting Notes AI, P&A said-vs-done, AiDE-drafted Ups & Downs, the reviewer brief, the risk radar, reports insights) call out to AiDE.

Environment variables

The server needs two: MCP_API_BASE_URL (the POD Governance API it calls) and MCP_API_TOKEN (your pgmcp_ token). MCP_PORT is only needed for the HTTP transport — stdio is the default and is what every desktop client below uses.

Claude Desktop setup

Edit claude_desktop_config.json — on Windows %APPDATA%\Claude\, on macOS ~/Library/Application Support/Claude/, on Linux ~/.config/Claude/:

{
  "mcpServers": {
    "pod-governance": {
      "command": "node",
      "args": ["<path>/src/mcp/dist/index.js"],
      "env": {
        "MCP_API_BASE_URL": "https://vishnuvm.centralindia.cloudapp.azure.com/PODGovernance/api/v1",
        "MCP_API_TOKEN": "pgmcp_YOUR_TOKEN_HERE"
      }
    }
  }
}

Claude Code CLI setup

claude mcp add pod-governance node ./src/mcp/dist/index.js \
  --env MCP_API_BASE_URL=https://vishnuvm.centralindia.cloudapp.azure.com/PODGovernance/api/v1 \
  --env MCP_API_TOKEN=pgmcp_YOUR_TOKEN_HERE

The server name carries the environment when you are not on production (pod-governance-staging, -qa, -dev), so several environments can be registered side by side without clashing. The Claude Code VS Code extension and the claude command line share the same registration.

AiDE Plugin — Remote Servers UI (recommended, no build)

In the AiDE sidebar click MCP Servers → the Remote Servers tab → Add a remote MCP server. Give it the server name, your environment’s /mcp URL, your pgmcp_… token as the Access Token, and leave the transport as Streamable HTTP. Every tool works over the hosted endpoint; only progress notifications are degraded, because the hosting buffers responses.

AiDE Plugin — stdio (.vscode/mcp.json)

To run the server locally instead, build it once (cd src/mcp && npm install && npm run build) and add it to .vscode/mcp.json in the workspace with "type": "stdio", node and the two environment variables. Everyone who opens the workspace then gets it.

HTTP transport

For a client that cannot spawn a process, run the server over HTTP instead:

MCP_PORT=3100 MCP_API_BASE_URL=https://vishnuvm.centralindia.cloudapp.azure.com/PODGovernance/api/v1 MCP_API_TOKEN=pgmcp_… node dist/index.js

What the server exposes

121 tools, 3 resources and 3 pre-built prompts — listing, forwarding and returning cycles, full CRUD on Ups & Downs and P&A items, submitting and reviewing P&A, reports, AiDE-suggestion drafting, admin CRUD for projects/PODs/users, Meeting Notes AI and Project Timeline. Long-running calls report progress notifications rather than appearing to hang.

Token management & the interaction log

A token acts as you and inherits your role and project access, so scope it read-only unless the assistant genuinely needs to submit or forward. Revoke instantly from My Profile if it leaks; every call it makes is recorded per-token under Admin → MCP Interaction Log, with slow-call flags. When you generate one you choose its expiry — 30, 60 or 90 days, or Never — and 90 days is the default; on that date it simply stops working.

Remove or reset the integration

Delete the server from your client (its config block, claude mcp remove, .vscode/mcp.json, or the Remote Servers panel), then revoke the token at My Profile → AI & Integrations → MCP Access Tokens → Revoke (admins revoke service tokens from Admin → MCP → Tokens). To start clean, revoke the old one, generate a fresh one, and add the server again. If you only need a new secret, Rotate on your own active token gives it one and keeps its name, scope, expiry and history; the old secret stops working at once. Edit changes a token’s name or expiry, or narrows it to read only.

👤 Profile & Tokens

The My Profile screen lets you manage all personal settings: your display name and account details, appearance (light/dark/system), email notification preferences, MCP access tokens, and active sessions.

Display name

Your display name is what appears in audit entries, notifications, the timeline feed and review screens — so it is not only cosmetic. Edit it and click Save; changes take effect immediately.

Download a copy of your data

On My Profile, Download my data (in Account Details) gives you one file with your profile and settings, your Ups & Downs and Priorities & Achievements entries, your action items, recurring tasks and focus sessions, the meetings and meeting series you own or joined, your project memberships and report schedules, the names of files you uploaded, and a record of your activity. Meeting transcripts and minutes are not included, because they hold other people's words too. Passwords, access keys and other secrets are never included.

The My AiDE runs card lists the AiDE runs you started, whether each worked, and when. Another card lists the name corrections you saved with remember these names when correcting a meeting, so you can change or delete them.

Switching your active role

If you hold more than one role — say both PM and Sales SPOC — the Active Role dropdown switches between them. The sidebar, the access rules and every screen adapt immediately, and your role badge in the top bar follows. Audit entries record the role that was active at the time of each action, so switching is part of the record, not a display preference.

My On-Time Record

The My On-Time Record card shows the share of your P&A weeks and Ups & Downs cycles you submitted on time, since you started — the same figures your reminder emails carry. Green at 90% and above, amber from 70%, red below that, with the on-time, late and missed counts and how your latest submission moved the figure. Only you can see your own record.

Appearance — light & dark mode

Pick Light (default), Dark, or System, which follows your operating system. The choice applies at once and is remembered per device. You can also toggle it from the command palette (Ctrl/⌘ K → Toggle dark mode).

Notification preferences

These control email only — the in-app 🔔 bell always shows everything regardless. Each switch covers a whole workflow: turning off Ups & Downs workflow or Priorities & Achievements silences its reminders and its approval / return notices together, and there is no separate switch for one without the other. Each switch saves as soon as you change it. The digest card underneath sets its frequency (Weekly, Daily or Off), what it includes, and can preview or test-send it.

From an email, without signing in. Every notification email has links at the foot that open an Email Preferences page for your address only, with Stop these emails, Stop all notification emails, a pause for 1 day, 1 week or 30 days, the same switches, and the digest frequency. Opening the link changes nothing until you click; sign-in, invitation and account emails always arrive.

Active sessions

Lists every device currently signed in — browser and OS, IP, last-active time and when the session started — with your own marked This device. Sign out any you do not recognise. In a normal browser tab sessions also end after 30 minutes of inactivity, with a 5-minute countdown warning; in the installed app that idle window extends to 24 hours.

MCP Access Tokens

Create and revoke Personal Access Tokens for the MCP server. Give each one a label naming the machine it lives on (“Claude Desktop — Work Laptop”) so revoking the right one later is obvious. The token is displayed once, in full — copy it immediately; only a hash is stored. You can also tick Limit to projects: the token then works only on those projects’ cycles, meetings, timelines and health, and is refused for anything else. See MCP Integration or the MCP Setup Guide link in the sidebar for the setup snippets.

Edit changes a token’s name or expiry, or narrows it to read only, and Rotate gives it a new secret, both without losing the token and its history.

⚙️ Admin Portal

The Admin Portal is available to System Admins and PMO Associates. It provides management screens for the entire platform. Everything about people and structure lives in one place, 👥 People & Organisation, as tabs (People & Roles, PODs, Projects, Hierarchy and more); the other areas each have their own sidebar entry.

Section	What you can do
Projects	Create, edit, and deactivate projects; assign all 6 SPOC roles
PODs	Create and manage PODs; assign POD Owners and Cluster Heads
Users & Roles	Invite users, change roles, deactivate accounts, multi-role assignments, and 🚪 Mark as left: hand a leaver's approval seats, meetings and open action items to a successor; work left with no owner stays in the warning panel, where 🤝 Hand over gives it to someone later; and, for a System Admin, 🕶️ Anonymise a deactivated person: their name and email become a pseudonym and the text they wrote is removed, while the audit log and meeting minutes stay; it cannot be undone
Org Hierarchy	Visual tree of the organisation, built from each project’s people, with expand/collapse
Scheduler	Manage recurring cycles, global automation settings, view run history
Email Templates	Edit, enable/disable, and test-send any system email template
Reports & Exports	On-demand CSV exports for cycles, P&A, users, returns, audit logs
Audit Log	App-wide append-only audit trail with CSV export
MCP Tokens	Manage all PATs across the platform; view expiry and revoke
MCP Interaction Log	Per-token activity log with slow-call flags and search
Meetings register	Every meeting on the platform, as a list or a calendar. Click a meeting title in the list, or a meeting chip in the calendar, to go straight to that occurrence — the meeting page still applies its own access rules, so this never opens a meeting you could not already see. Nudge owner and Export CSV per row.

Users & Roles

Go to Admin → People & Organisation and stay on the People & Roles tab. ✉️ Invite Users opens on the Single tab: enter a work email and role, optionally projects, and press Send Invitation. The link lasts the time set in Admin → Scheduler → Configuration → Global Automation Settings → Invitation Link Expiry (hours), 72 hours unless changed; if it is still unused after 48 hours a reminder goes with a fresh link, keeping that time or giving a full day if less is left. The pencil edits a user’s roles, projects and status; ⏸ deactivates and ▶️ reactivates. A user marked Invited has not signed in yet, and the resend icon issues a fresh link. Narrow the list with All Roles, All Statuses and the search box.

An invitation’s status says how long it has left, such as Expires in 2 days, amber under a day and red once it has expired.

The list shows when each person last signed in, with how long ago; an account that has never signed in says so. Filter by last sign-in finds accounts with No sign-in recorded or No sign-in in 90+ days.

🚪 Mark as left shows what the person holds — approval seats, meetings they own, open action items. Seats go to the next person up the chain; choose who takes the meetings and items in Hand these to (it suggests the person taking the seats), or Nobody, which leaves them in a warning panel where 🤝 Hand over gives them to someone later. 🕶️ Anonymise (System Admins, deactivated users only) replaces the name and email with a pseudonym and removes the text they wrote; the audit log and minutes are kept. It cannot be undone.

Bulk inviting users

In ✉️ Invite Users, switch to Bulk CSV and paste one email, role per line (up to 200; a header line is ignored; rows without a role take the default). Send Invitations processes each row on its own and shows Invited, Skipped or Error per email, so one bad row never stops the rest.

Projects

Admin → People & Organisation → Projects. + Add Project takes the name and the role slots (only the PM and Project Lead are needed to start a cycle). The pencil edits; pause stops new cycles; archive retires it with its history kept; ☑ Select pauses several at once. Project Contributors are people who take part but are not in the review chain — they appear in meeting rosters and never in the Ups & Downs approval chain.

Move to POD in the selection bar moves the selected projects to another POD together; each takes that POD’s Owner and Cluster Head as reviewers. The Chain column and the Chain needs attention filter show which projects’ cycles will stop at an empty seat or skip a review.

PODs

Admin → PODs: + Add POD with a POD Owner; projects join it from Admin → Projects. Existing projects keep their reviewers until you save them. Changing a POD’s Owner or Cluster Head does not change the projects already in it; Projects flags those still on the previous person, and saving one moves it (and its cycles in progress) to the POD’s current reviewers.

Scheduler

Admin → Scheduler opens cycles and fires reminders on its own once set up. + Add Cycle sets the name, frequency, start day and time, and projects; Run Now starts one at once; Pause and Archive stop it. The Activity & Run Log tab shows every run and any errors. Its jobs include the weekly leadership digest (see Ups & Downs → AiDE Risk Radar), off by default.

Some jobs watch for trouble so you do not have to open a screen to find it. Every day, System Admins and the PMO get one email listing the Ups & Downs cycles nobody can move, with the step each waits on and why; a cycle that stays stuck is listed again at most once a week. Admins are told when a tool used by AI assistants fails on at least one call in five over an hour, and when the clean-up of old scheduler runs falls behind.

On the ⚙️ Configuration tab, the 🏷️ Priorities & Achievements card sets the Default POD for unmapped submitters. A P&A item must name the project or POD it was for, and someone on no project can reach no POD, so this POD is offered to exactly those people. It shows Not configured until you pick one; the card has its own Save Changes.

The 🛑 Stopped Triggers panel names any job that used to run and has since missed whole scheduled runs, with when it last ran, when it should have run and how many runs it missed. Jobs that record a run only when they have work are listed apart, under 😴 Quiet: a gap there is not a sign the job stopped. Jobs that have never run yet are named in a short line below and are not treated as a fault. If the check cannot be done, the panel says Could not check rather than showing all clear.

Email Templates

Admin → Email, Templates tab (the Delivery Log tab beside it records what went out). Pick a template to see its subject, body and send counts; Edit changes it using {{variableName}} tokens; Test Send sends it to you with sample data; turning Enabled off stops it being sent; + New Template adds one, which sends only once enabled.

Subjects name a project by its short code, for example APSB rather than APS Bank, so more of the subject fits in an inbox; the token is {{projectCode}}. The email itself uses the full name, and a project with no code yet keeps its full name in the subject.

A few emails, such as the health digests and the weekly timeline reminder, build their body when they are sent because it holds a list or a table. Those show This body is built when the email is sent: a change to the body is saved but never reaches anyone, while the subject and the button are used as normal.

Org Hierarchy

Admin → People & Organisation → Hierarchy draws the organisation from each project’s people and roles. Check it before a cycle opens: gaps, such as a Project Lead with no PM, are shown in amber.

Seat Coverage

Admin → People & Organisation → Seat Coverage shows which senior seats are unfilled, on which active projects, worst-staffed first. The seats are Delivery SPOC, Sales SPOC, Sales SPOC (co), POD Owner and Cluster Head; fill an empty one from Projects. Only ACTIVE projects are counted. An unfilled seat is not merely cosmetic. The approval chain runs through these seats, so a gap becomes a review nobody can forward.

Remind on a project emails the person who can fill its empty senior seat: the POD Owner, or the Cluster Head when that is the gap. It sends at most once a day per project.

Access Requests

People ask for access from the Projects screen, instead of finding an admin by email; requests wait in Admin → People & Organisation → Access Requests, and the admin team is emailed. The requester types the project, since they cannot see the list. Grant the access in People & Roles first, then press Mark granted & tell them (marking alone adds nobody); a decline needs a note. Answered lists past decisions.

Reports & Exports

Admin → Reports & Exports gives System Admins and PMO Associates the same dashboard as Reports, plus scheduled reports and the 📥 On-Demand Exports panel: Cycle Status Report, Ups & Downs Trend, Priorities & Achievements Completion Report, User Activity Report, Return Rate Report and Full Audit Log.

Audit Log

Admin → Audit Log is every recorded action across the platform. Filter by entity, actor, action, role and dates, then Export CSV. The Audit Log topic explains how to read an entry.

The export holds at most 5,000 rows and now says when it was cut short. Tick Live to refresh the list every 30 seconds.

Email Log

Admin → Email → Delivery Log records every email sent, with its rendered body. For a failed send, filter Status: Failed, open the row to read why, then Resend. Click a recipient to see everything sent to them. Sign-in codes and invitation links are stored as [REDACTED] and cannot be resent from here; use Resend Invitation or Send Magic Link instead.

MCP Access Tokens (Admin view)

Admin → MCP → Tokens lists every token in the organisation with its owner, label, created and last-used dates. Revoke stops one at once; it cannot be undone, so check with the owner first unless it is a security incident.

A service account’s Edit renames it or changes the role it acts as; its tokens keep working.

MCP Interaction Log

Admin → MCP → Activity Log records every MCP tool call made by anyone’s AI assistant: who, which tool, how long, and whether it failed. Use it to audit AI use and spot unexpected calls; it exports to CSV.

Above the log, a summary table gives each tool’s calls, error rate, 4xx and 5xx counts, and its average, median, p95 and p99 times. Filter the log by Identity, Tool, Status or a From date, or use the quick search for an identity, tool or endpoint.

Webhooks

Admin → Webhooks tells other systems when something happens (a cycle approved, a submission returned, a meeting processed). ➕ New Webhook takes a name and URL; the signing secret is shown once, so copy it. The delivery table records every attempt and its response code (newest 200; Load more brings older ones). Webhooks are optional; the usual use is a team chat channel.

A failed delivery has ↺ Retry, which sends that event again. Each row shows its status (Active, Paused or Revoked) and how many of its events were delivered in the last 7 days, as a percentage coloured green, amber or red. Edit changes the name and the events and keeps the signing secret; turning Enabled off pauses the webhook, so it receives nothing until you turn it back on. The URL cannot be changed, because its secret was given to that endpoint: create a new webhook for a new address. Send test sends a signed test event to the URL now. Revoke stops it for good; the row stays in the list, marked Revoked, as a record.

Telemetry

Admin → Telemetry answers how fast, how often, and what went wrong, measured from real requests. Pick a window; tick Compare with previous period to see each headline figure against the window before. Slowest operations are ranked by the 95th percentile, because an average hides the slow tail people notice. Errors link to their traces, and a trace id can be looked up to follow one request end to end.

The page opens with four headline figures: operations, the share that failed, the slowest typical time and the number of services, and a Collection status card says what is really switched on here. In Slowest operations each row has a bar against the slowest; click an operation to see its own timings, a chart of how long its calls took, and its slowest calls, each of which opens its full request. A table gives the failure rate for each operation. A looked-up request reads as a timeline: each step is a bar where it started and as long as it took, coloured by service, and a failed step stands out.

Meetings register

Admin → Meeting Notes AI is an oversight register of every meeting in the organisation. It deliberately shows you metadata, not content: counts and status, not the transcripts or minutes. The tiles count meetings processed, open and overdue actions, failed, and stuck (started and never finished, the one to chase). A bar shows AiDE token adoption, with a list of who has not connected one.

Search and filter, switch between list and calendar, and click a meeting to open it; the meeting page applies its own access rules, so if you are not its owner or a participant it says the meeting was not found, or you do not have access to it. Nudge owner emails an overdue reminder (once an hour per meeting); Export CSV downloads the register honouring the filters you have applied. The reminder schedule is not configured here — that is Admin → Scheduler, and the wording is Admin → Email → Templates.

AiDE administration

Admin → AI Integration → AiDE shows what it is configured to be and every run it has performed. The Config tab shows the model, endpoint and timeouts, read-only on purpose (changing them is a release), and usage by run type with the median and 95th percentile. Runs lists every AiDE run not attached to a meeting, filterable by type, outcome and person, and exports to CSV.

A By type card above Runs gives calls, failures and timings for each kind of run; click a type to see only those runs. The Project filter shows one project’s runs. The Usage tab shows runs per day and who made them, heaviest first, over the last 7, 30 or 90 days. Test AiDE connection checks that AiDE answers, using your own saved token, and says how long it took.

Settings

Admin → Settings is a read-only view of the configuration this deployment is really running with, which helps when Dev, Test and Production behave differently. Secrets are never shown, only whether each is set. If users report missing emails, press Test connection first: it checks the mail relay, which tells a delivery problem apart from a template or schedule problem.

Settings also shows how much space attachments use, and files that no longer belong to anything can be removed from there; that clean-up is the one action on the page.

Status page notices

The public Status page checks each service in Production and Development and shows the build each one is running, but cannot say that sign-in failed this morning or that maintenance is planned for Saturday. A PMO or Admin posts that in its Incidents & maintenance panel: + Post a notice, choose Incident or Planned maintenance, set the start (leave the end blank while it is ongoing) and press Post notice; later, Mark resolved or Mark complete when it is over. Everyone can read the notices without signing in; the panel shows the last two weeks.