The Dashboard is the page Owners, Admins and Senior Leaders see first when they sign in. It's in five parts, and each part answers one question:
- Governance set-up: where are we, and what's next?
- Needs you: what needs me?
- How it's going: is it working?
- Next 90 days: what's coming up?
- Background: why are we doing this?
Each fact sits in one place, so nothing shows twice: a step that has become a row in Needs you leaves Governance set-up's In progress list, and anything in Needs you or under way in set-up is left out of Next 90 days. Reports holds the detail behind every figure.
On a wide screen, Governance set-up runs across the top, Needs you and How it's going sit in the main column, and Next 90 days and Background sit beside them. On a phone or a narrow window, it's one column in that order.
At the topLink to this section
The organisation's name, its time zone and View reports. Every date on the page is in that time zone.
Governance set-up: where are we, and what's next?Link to this section
While any set-up step isn't done, the page opens with Governance set-up: a dial of the six steps the cycle on our website uses, one segment for each, with the set-up steps grouped under them. It shows how many set-up steps are done overall ("4 of 10 steps done"), and in a list beside the dial, each of the six names with how far its steps have got ("1 of 2 done", Done, or Cannot start yet). On a phone, only the segment holding the next step is named beside the dial, with its figure under the name.
| The six steps | The set-up steps in each |
|---|---|
| Understand | Add and organise your people · Run The AI State of Play · List the AI tools already known about (or Import the organisation's tool list) · Add the organisation's AI policy |
| Leadership direction | Describe the organisation and its purpose · Agree the leadership position (or Record the leadership position, if it's already in writing) |
| Decisions | Decide on each tool · Agree the organisation-wide policies |
| Communicate | Get the extension installed (Highly recommended) · Brief the team · Download the How We Use AI Report |
| Engagement | Choose who is asked about each tool · Start The AI Pulse · Set the hourly rate (Pro and the trial) |
| Review | Hold the first Review (on the dial, but not counted as a set-up step: the countdown below says when it's due) |
The segment holding the next step is filled in navy, with Next step and that step's name beside it. The name links to where the work is done, and Owners and Admins get Continue set-up. The first time a step is next, a sentence under it says why it matters; after that, it's in the full list. Steps already under way show under In progress, up to three of them, with how far each has got ("Run The AI State of Play · 34 of 60 answered · closes 24 October 2026").
Show the steps opens the whole list under the same six names. Every step has one of four statuses:
-
Not started
-
In progress, with how far it has got
-
Done, with its date where the record holds one, for example "Sent 4 November 2026"
-
Cannot start yet, with what it waits for, for example "After adding your people"
-
The six steps don't have to go in order. Get the extension installed can be done while The AI State of Play runs, so Communicate can be under way while Decisions isn't finished. The figure beside each of the six names says where things stand, so nothing else on the dial needs to.
-
The next step follows Governance set-up's order: Your organisation, People, then The AI State of Play, which takes longest. The leadership position can be started at any time, and the extension, your AI tools and your existing policy can all be done while The AI State of Play runs.
-
Check and publish Our AI policies, in Governance set-up, is where you read the policies as staff will, before anyone is told. The first brief after The AI State of Play closes opens by thanking people for their answers.
-
The Engagement steps: the first two open once the extension rollout has started, and the hourly rate can be set at any time. Start The AI Pulse takes you to its step on Measurement → How it works, where its first start emails everyone it asks who has an account, with the message you've checked.
-
Three steps are marked by hand in Governance set-up: Add and organise your people, when you select I've added everyone; What The AI State of Play found, when you select Decide on each tool (or Agree the policies, once no tool is waiting), which also opens that step; and, where IT installs the extension, Get the extension installed, when you select I've told IT, which also starts the rollout. Every other step is marked done by the organisation's own records.
-
Set the hourly rate is on Pro and the trial only. On Standard it isn't on the list.
Governance set-up, the page in the menu, is where the work is done (Quick start: your first session walks through it). The dashboard shows the same list one level up. Once every step of Governance set-up and Measurement set-up is done, the dial leaves the dashboard and Governance set-up keeps the record. The first Review isn't counted as one of those steps: from the day both set-ups are done, the next Review countdown in Next 90 days says when it's due.
Needs you: what needs me?Link to this section
What needs a decision or an action from you, as up to five rows. Your own pulses come first ("Answer 3 pulses", which opens your list in the AI Adoption Portal), then everything else, longest waiting first. Show 3 more (or however many) opens the rest, and Show fewer closes them again.
Each row has an icon for its kind, a heading that links to where it's decided, and one line of facts, such as "Waiting since 2 October 2026". A row has a button only when the action happens right there on the dashboard: Send a reminder, Add to the next Review or Download the report.
Where there's a choice to make, the arrow at the end of the row opens Recommended, with what we suggest, and Why, with the reason and what it's based on (the leadership position, the organisation's policies, The AI State of Play, the vendor's terms or common practice). The first time a row with a recommendation appears, it starts open; after that it starts closed, and the recommendation is also at the top of the item's own page. The recommendation is a starting point.
Two or more rows of the same kind become one, with a count: "Decide on 6 tools", "Answer 3 incident reports". Its arrow opens each one. Renewals are never grouped, because each has its own date and cost.
A row that arrived since your last visit is marked New.
With nothing waiting, it says "Nothing needs you right now." While Governance set-up is showing and nothing is waiting, Needs you is left off the page, because the dial already shows the next step.
What puts a row hereLink to this section
| Row | When it appears |
|---|---|
| A tool waiting for a decision | 14 days after it was added or found |
| A renewal with no decision | Within 30 days of the renewal date |
| A decision past its review date | Once the date has passed |
| A request to use a tool | Waiting more than 5 working days |
| An incident report or a suggested change | Waiting more than 5 working days |
| A question people are asking | Asked 3 or more times, or open more than 7 days, and no published policy or tool decision answers it |
| People still to read the policies | Fewer than 80% have confirmed the policies that apply to them, 14 days after the current version was published or the team was briefed, whichever is later |
| The AI State of Play falling behind | Under 50% answered, with 3 days left |
| The How We Use AI Report out of date | 3 or more changes since the last download, or 90 days |
| The response rate falling | Once measuring: more than 10 points down on last month, unless a pulse started, or dropped out of the last 30 days, in the past month (that alone can move the rate) |
| A tool not being used | Once measuring: under 40% of its seats used, two months running |
| Satisfaction falling | Once measuring: down 5 points or more, with at least 5 ratings in both periods |
| The Leadership Position tool | When the round needs something from you |
| Findings and opportunities | While they wait on a decision |
Not every row has a recommendation. Incident reports, suggested changes, findings and opportunities, for example, say what needs attention and link to where it's handled.
Link to this section
- Send a reminder on The AI State of Play's row sends each person who hasn't answered their link again.
- Send a reminder on the policies row emails everyone with an account who still needs to confirm the policies that apply to them, at most once a day. The Policies and questions report has the same button. Neither goes to anyone who has already answered or confirmed.
- Send a reminder on the response-rate row is below.
- Add to the next Review records a finding about the tool, which the next Review lists.
- Download the report downloads the How We Use AI Report as it stands today.
You see how many people a reminder reached, never who.
Send a reminder on the response-rate rowLink to this section
When the response rate is falling, Owners and Admins can send the reminder email about waiting pulses now, rather than waiting for each person's usual reminder day.
- Who it reaches: everyone with a pulse waiting, whatever their usual reminder day and whatever they've had this week. Their usual weekly email then skips them that week, so nobody gets two.
- Who it leaves out: anyone who has turned the reminder email off in their own Settings. It's the same email, so their choice holds.
- Before it sends: if some of the people it would email were already emailed about a pulse today, it says how many and asks you to Send or Cancel.
- How often: once a calendar day for the organisation, from midnight in its time zone. Pressed again the same day, it says "A reminder went out today. You can send another tomorrow."
- When a press doesn't reach everyone (a send fails, or it runs out of time), it says so under the button, for example "Sent to 25 of 41 people. Send a reminder again to email the other 16." That press doesn't count toward the day, so you can press again straight away, and the second press emails only the people the first one missed.
- The record: the audit log notes who pressed it, with counts only.
If the rate stays low, Why is response rate low? goes through the usual causes.
How it's going: is it working?Link to this section
Up to three strips of figures: People, Governance and Results. Each figure has an icon, a label, its value, one line saying what the value is of, and its change on last month. A Behind badge appears at the top right only when the figure's row is in Needs you; a figure that's fine has no mark. A figure appears once it has something to count, never as a zero placeholder, and the whole section stays off the page until one does.
The question mark beside each strip's heading opens the definitions of its figures. "Change on last month" beside a heading names what every triangle in that strip is measured against: the last 30 days against the 30 days before, or for a count, today against the same count 30 days ago. An organisation under 30 days old, a trial included, has no last month yet, so no figure shows a change.
People: the four gaugesLink to this section
The People strip comes first, with each figure drawn as a gauge.
| Gauge | What it counts | Shows from |
|---|---|---|
| Active people | People who have signed in at least once, out of everyone on People who isn't deactivated ("92 of 119", "77% of all people") | Day one |
| People on the plan (or People on the trial) | Everyone on People who isn't deactivated, whether or not they've signed in, out of the most people at the current price, or the trial's or the account's limit ("119 of 140"). The mark on the arc is the active people ("92 active") | Day one. Owners and Admins only, and anyone looking at the demo |
| Response rate | For each pulse that ran in the last 30 days, the people who answered it out of the people it asked, added up across the pulses ("71%", "across 3 pulses"). The AI State of Play counts; the Leadership Position tool doesn't | The first pulse that runs |
| Extensions installed | Active people whose extension has connected in the last 30 days, out of all active people ("41 of 92", "45% of active people") | Once Get the extension installed has sent its invitations |
- Active people only goes up or stays flat. Once someone has signed in, they count. People drifting away shows up in the other two: Response rate falls when people stop answering, and Extensions installed falls when an extension goes 30 days without connecting.
- Over the limit, People on the plan's arc is full and the value says so ("152 of 140"). The banner on People says what happens next; Changing your plan or seats has the detail. Above 1,000 people, on Enterprise, or where no limit is recorded yet, it shows the count alone, with no arc.
- People on the plan's change has a soft purple triangle, because more people on the plan is neither good nor bad news.
- Extensions installed counts out of active people. Governance set-up's Roll out step counts out of everyone added, so the two can show different totals with the same number installed.
The 12-week linesLink to this section
Each gauge has a line under it showing the last 12 weeks: one point a week, ending on today's figure.
Most of what we measure is an estimate. Hours saved and satisfaction are what people tell us, and even a response rate moves with which pulses happen to be running. Any one figure can move for reasons that have nothing to do with how AI is going, so the direction over several weeks tells you more than any single number. That's why the gauges show a line.
The lines start the week after the first figures. We record the four figures once a night, and a line is drawn once it has two points: the first week's last recorded figure and today's. So the lines start on the Monday after the first night's figures. Until then, the strip's heading says when the lines will start ("Trend from 26 Oct"), and the figures have no line. We can't work out a past night's figures later, so a night that wasn't recorded stays a gap; we never fill one in. The change on last month for Active people and Extensions installed also waits until the nightly figures are 30 days old.
How to read a line:
- A point is the last figure recorded in that week (Monday to Sunday). The last stretch is the week so far, so on a Tuesday it covers two days, not seven.
- A break is a week with no figure to show: nothing was recorded, or on Response rate, fewer than 5 people answered or no pulse was running.
- A thin vertical line on Response rate marks a week when a pulse started, or dropped out of the last 30 days. Either can move the rate by many points with nobody answering any differently: a new pulse counts everyone it asks from its first day, while answers come in over the days after. The AI State of Play asks everyone, so the week it drops out can move the rate a long way.
- A thin stepped line on People on the plan is the limit, labelled "Up to 140".
There's no readout when you hover or tap. A screen reader reads every point of a line, and Reports has the detail.
GovernanceLink to this section
| Figure | What it counts | Shows from |
|---|---|---|
| Tools decided | Tools on the register with a decision ("9 of 11 tools") | Once Decide on each tool is done |
| Policies read | People with an account who have confirmed the policies that apply to them ("38 of 60 people") | Once Brief the team is done |
| Requests answered on time | Requests, suggestions and incident reports from the last 30 days answered within 5 working days ("7 of 8 in 30 days"), or "No requests" in a month with none | The first request, suggestion or incident report |
ResultsLink to this section
Once 5 people have answered the AI Pulse, over the last 30 days:
| Figure | What it counts |
|---|---|
| Seats used | Assigned seats where the person said they used the tool |
| Hours saved a person | Time people said the tools saved them, divided by the people who answered (Reported vs estimated hours says how to read it) |
| Satisfaction | How well the tools performed, across the tools with answers |
Before then, Results is one line saying how far the AI Pulse has got ("The AI Pulse · answered by 3 people"), with Assign each tool or Start The AI Pulse while either still needs doing. Owners and Admins can also Show the demo, to see what a measuring organisation's page looks like.
On Pro and the trial, a line under Results shows Estimated value of time saved (the hours people said they saved in the last 30 days, at the organisation's hourly rate) beside Tool costs, what the tools cost each month. Where only some tools have a cost, it says so ("Costs added for 6 of 9 tools"). With no hourly rate or no costs yet, it says "No rate or costs added", with Set the hourly rate or Add tool costs. How ROI is calculated explains the arithmetic.
When a figure is held backLink to this section
A figure from pulse answers needs answers from at least five different people. Below that, it reads "Not shown", with "Answers from fewer than 5 people" under it, and a change on last month shows only when both periods have answers from five. On a Senior Leader's dashboard, a figure that could point to someone in a small group also reads "Not shown", with "Hidden to protect a small group's answers" under it. The other three People gauges count records an Owner or Admin can already see on People, so they always show. In a smaller organisation, the Governance figures and the People gauges may be the only ones for a while.
The next stageLink to this section
On Standard, at a milestone (the first Review closing, or The AI State of Play closing with opportunities named), a row under How it's going says what the organisation's own numbers show. Its arrow opens what you get with Pro, with Compare plans (see Plans compared). Hide hides it for you, and the next milestone brings its own.
Next 90 days: what's coming up?Link to this section
The next Review countdownLink to this section
Once Governance set-up and Measurement set-up are both done, the top of Next 90 days counts down to the next Review: Next review due, the days to go ("26 days", "Tomorrow" or "Today") and the date ("13 November"). The first Review is due one cadence after both set-ups are done, counting from whichever finished later; after that, the date is the last Review's close plus the cadence. The cadence is the organisation's, monthly or quarterly, which you choose on Reviews. While a Review is open, the first one included, it reads Review closes and counts down to that Review's close date instead.
- Days are whole calendar days in the organisation's time zone. "Tomorrow" means the next calendar day there, whatever the time, and the date itself is "Today" until midnight.
- Once the date has passed without a Review being opened, it reads "3 days overdue", with a Behind badge. An open Review that hasn't been closed by its close date reads the same way, because a Review stays open until whoever runs it closes it.
- Owners and Admins get Open Reviews, or Open the Review while one is open. Senior Leaders see the countdown without a link, because Reviews is an admin page.
- While either set-up still has steps to do, and no Review has been opened, there's no date yet, so there's no countdown. The set-up dial shows instead.
Running a Review covers what happens in one.
The dated rowsLink to this section
Under the countdown, what's coming up in the next 90 days, soonest first: the date, an icon, the thing, and what happens.
| Row | What it shows |
|---|---|
| A tool renewing | A renewal in the next 90 days that isn't already in Needs you, with its seats and yearly cost where they're recorded |
| Decisions | Decisions whose review date falls in the next 90 days, on the earliest date ("2 due for review") |
| The leadership position | Its review date, for people who can read the record, while the Leadership Position tool isn't asking for something |
| The AI State of Play | When it's open again after set-up, with how many have answered |
Five rows show, then Show 2 more (or however many). Nothing in the list is in the past, and anything already in Needs you or under way in Governance set-up is left out. A row that arrived since your last visit is marked New.
Background: why are we doing this?Link to this section
Two rows, both closed until you open them:
- Why we are doing this: the organisation's purpose. Owners and Admins can Write the purpose (it reads "Not written" until someone does) or Edit it. Leaders can read it.
- Industry AI briefing: the month's briefing on AI in the organisation's industry, once there is one, with the industry and month beside it.
What each role seesLink to this section
| Who | What they see |
|---|---|
| Owners and Admins | Everything on this page, with every button |
| Senior Leaders | The same page, read only. Governance set-up with no Continue set-up and no step links. Needs you with only their own asks: their pulses, and anything the Leadership Position tool needs from them. How it's going for the whole organisation, without People on the plan, and with Response rate and Results leaving out the groups they lead, as their Reports do (a sentence under each says so whenever that leaves someone out, and Response rate then has no 12-week line). Next 90 days with no renewals, links only to pages they can open, and the countdown without a link. Background, read only |
| Group Leaders | Needs you with their pulses, and Response rate, Seats used and Satisfaction for the groups they lead, with no 12-week line. "Nothing to show yet." until someone in their groups has answered. Background, read only. No Governance set-up, no People strip and no Next 90 days |
| Members | The AI Adoption Portal instead of the dashboard |
Owners and Admins can also look at a demo organisation (Show the demo). There the page is read only: the links work, but nothing sends, downloads or continues set-up, and nothing is marked New.
What's not on the dashboardLink to this section
No comments, no single answers and no names. The dashboard shows figures for the organisation and its groups. Reports has the rest: the Governance reports (The AI State of Play results, the leadership position, Tools and renewals, Policies and questions, Requests, Reviews and the How We Use AI Report), ROI, the comment themes, the comparisons and the AI Impact Report.
If you'd like someone to help turn these numbers into decisions, see Getting outside help.
Something missing or out of date here? Tell us.