Meeting Input is the central data entry point. Paste any raw text - meeting minutes, status emails, site visit reports, change request documents - and an AI engine reads it and populates the relevant sections automatically. You review everything before it touches your data.
| Category | Trigger phrases (examples) | Goes into |
|---|---|---|
| Decisions | Any clear statement of a decision made | Decision Log (09) |
| Actions | Owner + task + deadline combinations | Decision Log - standalone |
| Risks | "risk", "might", "concern", "exposure" | Risk Register (08) |
| Budget - allocated | "budget of", "approved at", "cost plan" | Budget (04) - Allocated |
| Budget - spent | "we have spent", "actual to date", "invoiced" | Budget (04) - Spent |
| Budget - forecast | "expecting to spend", "projected cost" | Budget (04) - Forecast |
| Schedule impacts | Date slippage, revised date, new deliverable | Schedule (05) & Timeline (07) |
| Issues | "problem", "issue", "defect", "blocker" | Issue Tracker (11) |
| Change requests | CR number, variation order, scope change | Change Log (10) |
| Assumptions | "assuming", "we assume", "if X happens" | Assumptions Log (03) |
| Dependencies | "blocked by", "can't start until", "waiting for" | Dependency Tracker (12) |
"3 weekends of validation starting 5 July" produces three separate Schedule entries with individual dates - not one collapsed line.
After extraction, every item is shown with a checkbox. All are selected by default. Deselect anything that isn't relevant before committing.
The AI's judgement on decision versus change versus risk is usually right and occasionally not. Each row in the review panel carries an Add as dropdown listing all nine destinations. Change it and the item is filed under the new type when you commit - a decision becomes a change request, a change becomes a risk, and so on.
Nothing is written to your registers until you press Add selected, so changing the dropdown costs nothing. If you spot a misfiled item after committing, use Undo last import and re-run.
Every extraction is tagged with a batch ID. The Undo button removes all items added in the most recent extraction - and only those. Items added manually, or from earlier extractions, are untouched.
If a field isn't present in the notes, it's left blank. The extraction only produces items actually supported by the text.
The deck ships dark. If a full day of it is hard on the eyes, the ☼ LIGHT button under the guide links in the sidebar switches to a light palette. On a phone it sits at the right-hand end of the top bar.
The setting is shared with the rest of Operational Mind, so switching here also switches the home page and the document chat, and it is remembered on that device. If you have never chosen, the deck follows whatever your operating system is set to.
Nothing else changes: the same data, the same layout, the same RAG colours. Exports are unaffected. PowerPoint decks, Excel registers and Word reports always use the printed brand palette regardless of which mode you are in.
The top-level view across all your projects simultaneously. Pulls live data from every project and presents it in one place - no project-switching needed.
The command centre switches context and takes you to the Project RAG Summary for that project.
A headline strip sits above everything and gives the five-second answer: how many projects, whether any are red, how many actions are overdue, how many changes await approval. When there is nothing pressing it says so, rather than showing a row of zeroes.
Below it the page is grouped into six labelled bands, ordered by how often you act on them rather than by importance:
| Band | Holds |
|---|---|
| Health | Project status, open risks and issues, as donuts |
| Needs attention | Actions overdue, blocked or due this week; highest-priority risks; decisions and pending changes |
| People | Workload by owner |
| Trends | Risks and issues over time, budget allocated against spent |
| Projects | The full project list |
| Settings | Data, backup and sync; currency |
Trends sits below People and Needs attention on purpose. You act on an overdue action daily and read a trend line weekly.
Click any band heading to fold it away. A collapsed heading always keeps its summary - 2 actions · 1 open risk, or 1 stretched - so folding a band can never hide something that needs you. Your choice is remembered on that device, like the light and dark setting, and is not synced.
A filter bar sits under the heading on the Portfolio Dashboard. Filter by health (All, Red, Amber, Green) and by owner. A counter shows how many projects are visible and Clear resets it.
Health narrows the dashboard to projects at that RAG. Everything below recalculates.
Owner turns the dashboard into that person’s view. It narrows the projects to the ones they are on, and it narrows the items to theirs: the highest-priority risks list, both trend charts, the donut charts and the workload card all show only their work.
| Card | With an owner selected |
|---|---|
| Highest-priority risks | Their risks only |
| Open risks / issues over time | Their items only |
| Workload | Their row only, retitled with their name |
| Projects list | The projects they are on, keeping project-level budget and RAG |
| Budget: allocated vs spent | Hidden. See below |
Budget belongs to a project, not a person. Showing project spend underneath somebody’s name would invite a reader to take it as their spend, which is the sort of thing that goes badly in a resourcing meeting. The card explains itself and comes back when you clear the owner filter. The projects list keeps project-level budget and RAG for the same reason: those figures are the project’s, and are labelled as context rather than as that person’s numbers.
Two line charts show how many risks and issues were open at the end of each month. These are reconstructed from the dates already stored on every item, so they show real history going back to the first item you ever logged, up to twelve months.
This one works differently and it is worth knowing why. Budget figures are overwritten each time you edit them, so there was never a record of what they used to be. From now on the Command Centre writes a small monthly snapshot of each project: its RAG, budget totals and open counts. The chart draws from those.
That means the budget chart starts empty and needs two months before it shows a line. It fills in on its own from there. Nothing is asked of you, and the snapshots back up and sync with everything else.
A cross-project view of who is carrying what: open risks, open issues, blocking issues, open actions and anything overdue, totalled across every project in view.
The bar on the right orders the list, and nothing more. It is drawn relative to the busiest person on screen, so a full bar means “most loaded here”, not “overloaded”. There is no target and no ceiling. Hover it for the arithmetic.
| Item | Counts as |
|---|---|
| Blocking issue | 5 |
| Overdue action | 4 |
| Open risk | 2 |
| Any other open issue | 1 |
| Any other open action | 1 |
So somebody holding one blocking issue and two open risks scores 5 + 2 + 2 = 9. The line above the bar gives the actual counts, which is the part worth reading in a resourcing conversation. The bar is red if they hold a blocking issue, amber if something of theirs is overdue.
Anyone holding a blocking issue in more than one project is flagged stretched, which is the bottleneck worth catching early. Click a name to switch the dashboard into that person’s view, as described under Filters above.
A cross-project list of the work that needs a person, rather than every open task. Three things qualify:
| State | Meaning |
|---|---|
| Overdue | Past its deadline and not Done. Shown in red with how many days late |
| Blocked | Has a blocker recorded. No deadline needed, because these rot quietly |
| Due this week | Lands within seven days |
Sorted worst first: the most overdue at the top, then anything blocked, then whatever is due soonest. An action that is both late and blocked is listed as overdue, since the date is the harder fact. Capped at the eight most pressing, with a count of the rest.
The card only appears once you have open actions somewhere, and it says so plainly when nothing is overdue, blocked or due this week.
Three things can be done without leaving the dashboard:
All of these write to the project that owns the item, wherever it lives, not to whichever project you happen to have open.
The at-a-glance health check for the current project. Quality now calculates itself from the three measured lamps, and you can override it whenever your judgement differs from the arithmetic.
The Quality card offers four buttons: Auto, Green, Amber and Red.
Overall RAG is still the worst of all four values. Setting Quality to Green on a project running 20% over budget gives you a green Quality lamp and a red Overall. Escalation works the other way: setting Quality to Red forces Overall red regardless.
They had no Auto setting, so they read as Auto and will show a calculated Quality on next load. A project you had deliberately set to Green while its budget was Red now shows Red. One click on the colour you want puts it back under manual control.
These were undocumented, which made the odd result look like a bug rather than a rule. Each lamp reads the registers directly.
| Lamp | Red | Amber | Green |
|---|---|---|---|
| Schedule | Worst slip over 14 days | Worst slip of 4 to 14 days | No slip, or 3 days or fewer |
| Budget | Spend over 10% above allocation | Spend 2 to 10% above | Within 2% |
| Risk | Any open risk at High likelihood | Three or more open risks | Fewer than three, none High |
| Quality | The worst of the other three, unless set by hand | ||
The three-day schedule tolerance is deliberate. A lamp that turned amber the moment any date moved by a day would be amber permanently and tell you nothing. Because of it you can legitimately see GREEN above 1 slipped, so the lamp now spells out the tolerance: 1 slipped, worst 2d - within 3d tolerance.
Risk is deliberately strict. One open High-likelihood risk turns the lamp red, matching the red pill on that risk exactly. If the register shows a red risk, the lamp is red. The two can no longer disagree.
This page is built to be scanned in a weekly check-in, so it is grouped into three bands:
As on the Portfolio Dashboard, a collapsed heading keeps its counts, so 3 open risks · 2 decisions reads from the closed band. Nothing has been removed from the page; the long lists are one click away.
The PowerPoint export from this page was always the clean executive artefact, which leaves the page itself free to be the working view.
See How Meeting Input works above for the full breakdown. Supported input: meeting minutes, status emails, change request documents, site visit reports, or any plain text.
"Next Friday" or "end of next week" are converted to absolute dates based on today's date at the time of extraction.
A register of things the project is assuming to be true. Assumptions that turn out to be wrong are a common source of unmanaged risk - this log makes them visible and actionable.
| Field | Description |
|---|---|
| Text | The assumption itself, stated plainly |
| Impact if wrong | What breaks - cost, schedule, scope - if the assumption turns out false |
| Owner | Who is responsible for validating this assumption |
| Last validated | When it was last confirmed to still hold |
| Status | Unknown · Valid · At Risk · Broken |
Click the status label to cycle. Items are grouped by status with Broken at the top.
Any assumption at At Risk or Broken status shows a ⚠ button. Clicking it creates a new High priority risk using the assumption text and impact - no re-typing required.
Trigger phrases: "assuming", "we assume", "this depends on", "if X happens", "assuming the vendor", "assuming we get approval".
Track budget by line item - typically one line per supplier, contract, or cost centre. Three amounts per line give a clear picture of where money stands.
| Field | Meaning |
|---|---|
| Allocated | What has been budgeted, approved, or contracted for this line |
| Spent | What has actually been paid or invoiced to date |
| Forecast | What is projected to be spent by completion |
Total rows at the bottom sum all three columns. Variance = Forecast minus Allocated. Edit any line with the ✏ button alongside the delete button.
The AI distinguishes amounts by language: "we have spent 1.6M" → Spent. "Budget approved at 2M" → Allocated. "Expecting to spend 500k" → Forecast. Only the relevant field is populated; others stay at zero. Figures are read as plain numbers in whatever currency your notes use, and shorthand like 2M or 850k is expanded to the full amount.
The milestone list view - the same data that powers the Gantt, in table form. Use this to scan dates quickly or add a new milestone with full detail.
| Field | Description |
|---|---|
| Name | Milestone or deliverable name |
| Planned date | The original baseline date |
| Forecast date | Current expected completion date |
| Actual date | Date completed (if done); triggers Completed status on the Gantt |
| Owner | Person or team responsible |
| Status | On Track · Slipping · At Risk · Completed · Blocked - or blank to auto-calculate |
| Blocker | What is preventing completion |
| Notes | Context, references, caveats |
A monthly calendar that aggregates key dates from across the project automatically. Navigate months using Prev / Next buttons.
Dates are read live from the source sections - no duplication needed.
The CW column down the left gives the ISO 8601 week number for each row, the same numbering used across most of European project planning. Hover a number to see the Monday that week begins on.
Click any day to open its detail. Events you added yourself carry a ✎ to edit and a ✕ to delete. Editing covers the title, date, type and notes, so rescheduling is a change of date rather than a delete and retype. Move an event to another month and the calendar follows it there.
Entries that come from elsewhere - milestone dates, risk and issue deadlines, decision dates - are shown but not editable here on purpose. They belong to their own registers, and two places to edit the same record is how the two stop agreeing.
A scrollable horizontal Gantt chart built from the same milestone data as the Schedule. The visual layer makes slippage and patterns immediately legible in a way a list cannot.
| Element | Meaning |
|---|---|
| Grey tick mark | Planned (baseline) date |
| Coloured bar | The span between planned and forecast, coloured by status |
| Coloured tick mark | Forecast date (right end of bar) |
| Gold diamond ◆ | Actual completion date (if recorded) |
| Variance label | e.g. "+12d", "3d early", "on time" |
| Status label | Right-aligned, coloured to match bar |
| Gold vertical line | Today |
Leave the Status field blank and the Gantt calculates it:
Manually set Completed or Blocked always overrides the calculation.
Click any milestone row to open a detail panel beneath the chart showing the full milestone details plus three live summaries from the rest of the project: open actions (not Done), open risks (not Closed), and at-risk dependencies (At Risk or Blocked). Click the row again or ✕ to close.
Four range buttons control the width of the view:
| Range | Shows | Best for |
|---|---|---|
| Fit | Everything, auto-scaled to fit the screen | The default. A quick look at the whole project |
| Month | A three month window at full detail | Delivery detail and near-term variance |
| Quarter | Twelve months | Annual planning and quarterly reviews |
| Year | Three years | Board packs and macro-level planning |
Tick labels coarsen automatically as the range widens: months become quarters, then years. When a view crosses a year boundary, a year band appears above the ticks. Variance labels are hidden on the widest ranges, where they would overlap.
← Prev and Next → step through time in proportion to the range: one month on Month, four on Quarter, twelve on Year. They are inactive on Fit, which already shows everything.
This project is the default. All projects puts every project in your portfolio onto the same axis, each in its own labelled lane, so you can see where milestones collide or where a dependency sits in someone else’s plan. Click any row to drill down; the detail panel names the project the milestone belongs to and pulls that project’s open actions, risks and dependencies.
Projects with no dated milestones are left out of the portfolio view rather than shown as empty lanes.
Every item in every register carries a 🔗 button next to the delete cross. It opens a small panel where you can attach as many web addresses as you like: the SharePoint folder, the Teams message, the vendor report, the ticket, the email thread.
Nothing is uploaded and no file is stored. Only the address is kept, alongside the rest of that item, so it backs up, restores and syncs exactly like any other field. Your documents stay where they already live, under your organisation’s own permissions.
This is deliberate. It means a blocking issue can point straight at its evidence without the Command Centre becoming a document store, and without anyone needing a parallel spreadsheet just to remember where things are.
Anything that is not http, https or mailto is refused. That rules out a local file path on your own machine, which would not open for anyone else anyway, and it closes off addresses that could run code when clicked.
Every register sheet in the Excel workbook now has a Links column carrying the label and address of each attached link, one per line. The workbook is as useful as the screen.
Every ✕ deletes straight away, with no confirmation to click through. A toast then sits in the corner for ten seconds offering Undo, and the item goes back exactly where it was in the list, not at the top.
This covers all thirteen registers plus budget lines, milestones and calendar events. It exists because a register that loses history silently is worse than useless for reporting, and a confirmation box you dismiss fifty times a day stops being read by the second day.
Once the ten seconds pass the item is gone for good. If you are unsure, the safer move is closing an item rather than deleting it: a closed risk keeps its full history, still exports, and can be sent to Lessons Learned.
A standard risk log with lifecycle tracking. Risks are listed in order of entry. Closed risks remain for audit purposes but are visually separated.
| Field | Description |
|---|---|
| Title | Short name for the risk |
| Impact | What happens if the risk materialises |
| Likelihood | High / Medium / Low |
| Owner | Who is managing mitigation |
| Mitigation | What is being done to reduce the risk |
| Deadline | When mitigation must be in place |
| Status | Open · In Progress · Closed |
Click the status text to cycle. The ✓ button jumps directly to Closed. Status text is colour-coded: blue for Open, amber for In Progress, green for Closed.
The AI populates Title, Impact, Likelihood, Owner, Mitigation, and Deadline where stated. Status defaults to Open.
The bar above the register searches title, impact, mitigation and owner at once, and filters by owner and by status. A counter shows how many of the total are showing. Typing does not disturb the cursor.
This page used to claim it was sorted by "impact x likelihood". It was not, and could not be: impact is free text with no score attached. Risks are grouped and sorted by likelihood, highest first, and each card now names the likelihood that put it where it is.
An open issue in the Issue Tracker carries a ⚠ that copies it here as a risk, at High likelihood if it is Blocking. See Section 11 for why that is a button rather than an automatic sync.
A record of project decisions with accountability built in. Decisions can have actions attached directly to them, keeping who-does-what-by-when linked to the reasoning behind it.
| Field | Description |
|---|---|
| Decision text | What was decided |
| Approved by | Who made or approved the decision |
| Date | When the decision was made |
| Cost impact | Financial implication if any, in your selected currency |
| Schedule impact | Time implication if any (e.g. "2 weeks") |
Each decision card has a + Action button. Actions added this way are linked to that decision. Fields: action text, owner, deadline, status (Not Started / In Progress / Done), % complete, blocker.
Actions from Meeting Input that aren't linked to a decision appear in a Standalone actions section at the bottom. Each has a Move button to link it to a decision and optionally update the owner and deadline at the same time.
When any action is past its deadline and not Done, a red banner appears at the top. When actions are due within 5 days, an amber banner appears. Both also surface on Section 01 (Project RAG Summary) and Section 00 (Portfolio Dashboard).
Overdue actions carry a ⚠ button. One click logs a High-likelihood risk carrying the action text, its owner, and either its blocker or the deadline it missed. An action that has slipped its date is a live threat to the plan, and this puts it in front of the people who read the risk register rather than leaving it in a list they may not.
This is the same escalate pattern the Assumptions Log already uses for broken assumptions. The action stays where it is; nothing is moved or deleted.
A formal register for change requests and variation orders. Keeps scope control visible and auditable throughout the project.
| Field | Description |
|---|---|
| CR ID | Change request number or reference (e.g. CR-012) |
| Title | Short description of the change |
| Category | Type of change - scope, design, regulatory, etc. |
| Description | What has changed and why |
| Cost impact | Additional or reduced cost, in your selected currency |
| Schedule impact | Time effect on the programme |
| Drawing refs | Specification or drawing numbers affected |
| Status | Pending · Approved · Rejected |
If notes include a CR number, variation order reference, or formal scope change, Meeting Input extracts all available fields including cost impact, schedule impact, and drawing references.
A running log of operational problems that need resolution. Distinct from risks (potential events) - issues are things that have already happened and need addressing.
| Field | Description |
|---|---|
| Title | Short name for the issue |
| Impact | What it is affecting - schedule, quality, cost, safety |
| Owner | Who is resolving it |
| Deadline | When it must be resolved |
| Blocking | Flag: checked if this issue is stopping other work cold |
| Status | Blocking · Open · Closed |
Open issues carry a ⚠ that creates a matching risk, at High likelihood if the issue is Blocking and Medium otherwise, carrying the title, impact, owner and deadline.
Why a button and not an automatic sync. An issue is a problem that already exists; a risk is one that might. Copying every blocking issue automatically would count the same problem twice, in two registers and again in the RAG roll-up, making the project look worse than it is. The button keeps the two registers connected and leaves the judgement where it belongs.
Tracks cross-workstream dependencies - where one piece of work cannot proceed until something else is delivered. Particularly important on multi-supplier or multi-team projects.
| Field | Description |
|---|---|
| Dependent | The work or workstream that is waiting |
| Blocker | What it is waiting on |
| Blocker owner | Who controls or is responsible for the blocker |
| Impact | What happens to the dependent work if the blocker is late |
| Resolution date | When the blocker is expected to be resolved |
| Workaround | Any interim approach if the blocker can't be resolved in time |
| Status | On Track · At Risk · Blocked |
Any dependency at At Risk or Blocked status appears in the Gantt drill-down panel (Section 07), giving a cross-section view of what is constraining each milestone.
Trigger phrases: "blocked by", "can't start until", "depends on", "waiting for", "critical path", "prerequisite".
The post-mortem log: what to carry into the next project, and what not to repeat. Most registers record what is happening now. This one is the only part of the Command Centre written for the project after this one.
| Field | Description |
|---|---|
| Title | Short summary of the lesson |
| Category | What went well · What went badly · Process · Technical · Communication |
| What happened | The situation, plainly |
| What we would do differently | The actual lesson, phrased for the next project |
| Raised by | Who logged it |
When you mark a risk or issue resolved, the dialogue offers to record it here as well. Tick the box and the title, impact and your resolution note carry across, tagged with where it came from. The knowledge in a closed risk outlives the closure, and this is the only place it survives.
Lessons take attached links like any other item, and appear as their own sheet in the Excel workbook.
On the Project RAG Summary page, the ⬇ PowerPoint button builds a branded status report: a title slide with the overall RAG, the four RAG lamps with key figures and top risks, recent decisions and overdue actions, a milestones table, and a drawn Gantt timeline with status-coloured bars.
On the Portfolio Dashboard, the ⬇ PowerPoint button exports the leadership view: a Portfolio RAG Summary (health, combined budget, open risks, needs attention, plus a colour-coded RAG grid for every project), the portfolio overview table, and the highest-priority risks and pending approvals across all projects.
On the Project RAG Summary page, the ⬇ Excel button downloads one workbook with a sheet per register: Risks, Decisions, Actions (with their linked decision), Issues, Changes, Assumptions, Dependencies, Milestones and Budget. Opens in Excel or Google Sheets; save any sheet as CSV from there if needed.
Every sheet arrives as a proper Excel Table: a styled header row, banded rows for readability, filter dropdowns on every column, the header frozen so it stays put as you scroll, and columns sized to their contents. Budget and cost figures come through as real numbers with thousands separators, so they total correctly rather than arriving as text.
Because each sheet is a genuine Excel Table rather than a plain grid, sorting, filtering and structured references all work straight away, and adding a row extends the formatting automatically.
On the Meeting Input page, after you extract a set of notes, a ⬇ Weekly Meeting Report button appears. It produces a branded, client-ready Word document: meeting date and times (or the time you generated it if none are stated), an attendees table with a blank Role column to complete by hand, a discussion summary, a plain-English meeting summary, an action items table with owners and status (Not Started / In Progress / Completed), the key discussion points and notes, and a conclusion.
Unlike the PowerPoint and Excel exports (which are built instantly from data you already have), the meeting report re-reads your notes as a meeting to draft the summaries and narrative. It counts as one against your extraction allowance, and takes a few seconds longer. It is written in British English. Always review it before sharing with a client.
Your data is stored in this browser by default, so protect it. On the Portfolio Dashboard, ⬇ Download backup saves every project as a single file. ⬆ Restore from backup loads that file on any device: projects with a matching ID are replaced by the backup version, everything else is added alongside.
Cloud sync (optional): the Enable cloud sync button on the same card keys your projects to your login, encrypted, so they follow you automatically between laptop and phone. It is off unless you switch it on, you can disable it anytime, and a one-click control deletes the cloud copy entirely. Documents you upload to the chat are never stored or synced either way.
The Currency card on the Portfolio Dashboard sets how amounts are displayed everywhere: budgets, cost impacts, the Portfolio Dashboard totals, the PowerPoint decks, the Excel workbook, and the labels on the entry forms. Pick from the dropdown, or choose Other and type any three-letter ISO code. The default is GBP.
The AI records amounts as plain numbers whatever symbol appears in your notes, so nothing written in a transcript can change your display currency. If it worked the other way, one meeting mentioning a cost in another currency could silently relabel an entire project.
The consequence: if a single set of notes mixes two currencies, every figure from it is shown in your selected currency. Enter those amounts by hand if the distinction matters.
With sync off there is no copy on any server, so download a backup regularly, the same way you would save an important spreadsheet. With sync on, your cloud copy has you covered.
All exports and backups are generated entirely in your browser.
During the beta period, usage is capped per browser. The limits are generous for normal use.
| Limit | Value |
|---|---|
| AI chat messages | 500 per session |
| Document uploads | 20 per session |
| Maximum file size | 10 MB per file |
The platform uses Clerk authentication. Your account must be on the approved list to access the Command Centre. If you see an "Access Denied" page, contact the team to check your account status.