Watch out for these icons.
This one means your answers are anonymous.
This one means they're not.
Every pulse shows one of them before its first question. When a pulse is anonymous, your name isn't shown with your answers to anyone in your organisation. In a small group, or from something you write in a comment, someone may still be able to work out an answer is yours. Identifiable pulses say so before you start, and show your name with your answers.
The AI State of PlayLink to this section
The AI State of Play is a survey your organisation asks everyone once, to find out how AI is used at work. It isn't anonymous: you may be identified from this survey, and its first screen says so. You can answer it from the link in your email, whether or not you've set up an account.
Its first screen also says who the link was sent to. If you haven't set up an account and it has someone else's name on it, choose Not for me, then Stop this link: the link stops working and your organisation is told, so it can check the address. Your answers are recorded only when you press Submit.
What your organisation seesLink to this section
On an anonymous pulse, your organisation sees the answers without names. They see averages, totals, trends and comments word for word (your team leader, for the groups they lead), and results need answers from at least five people before they appear. Some of them can also read single answers and the day each was given, with no name shown.
On an identifiable pulse, your organisation sees your name beside your answers, and so does your team leader. They can read identifiable answers side by side, by person.
Whatever the pulse, your organisation can also see whether your extension is installed.
Requests, reports and feedbackLink to this section
When you send something in, the people who look after requests in your organisation can see it's from you, so they can get back to you: a request to use something, a suggestion about a policy, something you found worth knowing, or a report of something that went wrong. They read each one and answer it.
On Pro, where your organisation has switched it on, your team leaders (the leaders of the groups you're in, and of the groups above them) can read your requests, suggestions and finds too, from the last 90 days, if you sent them while you were in that group. They can see who sent each one, and they can't answer them. The form tells you before you send: selecting the person icon beside the button says Your team leader can see this too, and you can tick Keep from my team leader. A report of something that went wrong only ever goes to the people who answer. Under Your requests, each item says who sees it, and your data download says who could read each one.
General Feedback is like a pulse: its first screen says whether it's anonymous or identifiable, and in your AI Adoption Portal, where your team leaders can read it, it says that too. They read it as they read a pulse's comments, and anonymous feedback shows to them only once at least five people in the group have given some, with the day it was given and without a name or a tool beside it.
Whatever you send, leave out health or personal details, and other people's names.
Can I see my own answers?Link to this section
Yes. Sign in at any time and your AI Adoption Portal shows your own answers, grouped by pulse, with each question as it was asked. Where your organisation runs pulses about its AI tools, it also shows how many you've answered and your hours-saved estimate. Answers you gave from a link in an email, before you had an account, appear there once you set up your account.
Who elseLink to this section
- Your colleagues can't see your answers. What they see when they sign in is their own, not anyone else's.
- Nobody outside your organisation can see them. Your answers are kept in your organisation's account. One service outside it reads written answers: to group them into themes, we send the words, without your name, to Anthropic, our AI provider.
- The GAiGE staff. We see what your organisation sees only when it asks us for support, and we don't build profiles of individuals.
When you're asked something just onceLink to this section
Some pulses are one-offs. Instead of coming round every week, they open, you answer once, and they close. Organisations use them to take a starting picture: what AI people already use, what would help most, what's getting in the way.
You're asked once, and then you're done. The moment your answer reaches us, that pulse stops being offered to you. That holds even if you reinstall your browser or move to a new laptop.
You'll be reminded, and the reminders get closer together. In the extension, if the organisation has set a date it needs answers by, the pulse comes back about weekly while that's more than ten days off, every few days inside ten, and daily in the last stretch. It never comes more than once a day. A one-off with no date sits on the middle setting, so nothing can ask you daily indefinitely. Separately, if something is waiting for you, we send one reminder email a week at most. You can turn that email off in Settings, under Account, though you may then miss things your organisation needs from you. The AI State of Play has reminders of its own: until you've answered, its link comes again by email on day 4, day 9 and the day before it closes, on working days, unless your organisation has switched them off. Your organisation can also send its link again before it closes.
The reminders come to you, and only to you. Nobody else is emailed about whether you've answered. On an anonymous pulse, no screen lists who has or hasn't answered.
A check-in, if you haven't answered for a while. If you haven't answered a pulse in three months, we may email you once to ask one question: is something getting in the way? You can choose a reason from a short list, or leave it. Your reply isn't linked to you: your organisation sees how many people we asked and how many chose each reason, never who. We send it at most once every six months, and not at all if you've turned reminder emails off.
You can always leave. Every screen offers a way out: in the extension, snooze it for a few hours, choose Later, or close it; in your AI Adoption Portal, choose Leave. Close a one-off half way through and it goes back in the queue rather than being used up, and it tells you so.
A question that needs an answer has no Skip this one button. Optional questions have one, and the open questions are always optional. You can always leave the whole pulse for later.
What your organisation sees at the endLink to this section
On an anonymous one-off, three things, and none of them is a name. On an identifiable one-off, your organisation also sees each person's answers with their name.
- How many people have answered. A percentage and a count ("26 of 42 people"). Not which 26.
- How each question was answered. Charts of the spread, and a 0-100 reading for the scale question.
- What people wrote. The open answers, grouped with other answers that said something similar, most common first. Each group can be opened to read the answers underneath it, word for word.
On an anonymous pulse, a question needs five answers before any of it appears. Below that the report says how many people have answered and shows nothing else. The same floor applies to teams: a team where fewer than five people have answered shows its name and stops there, and a team the pulse reaches fewer than five people in is left out of the list entirely.
Two things to knowLink to this section
Your words are read. On a one-off, the open answers are the whole point, so they will be read closely and quoted, exactly as you typed them. On an anonymous pulse your name isn't shown, and a heading in that report is always something a real person wrote rather than a summary we've invented. But write knowing it will be read, which is the same advice as the caveat below.
In a very small team, a count can point to a person. "Five of six have answered" doesn't say who the sixth is, but in a team of six it doesn't hide much either. Nothing about what you said is exposed by that number, but in a small group someone may still be able to work out an answer is yours.
One caveat: free-text commentsLink to this section
When you leave a free-text comment ("any other comments?") on an anonymous pulse, your name isn't shown with it. But if the comment itself identifies you, the words can give you away.
Example of a safe comment:
"Copilot's autocomplete is great for boilerplate but keeps suggesting deprecated APIs in our Python 3 codebase."
Example of a comment that identifies you:
"As the only person on the procurement team who uses Jasper, I find it frustrating…"
If staying unnamed matters to you, keep the comment about the work, not about your specific situation. When in doubt, err on the side of brevity.
What about negative or critical answers?Link to this section
Negative answers (low ratings, "I lost time because of it", blunt free text) are fine, and they're often the most useful: they show your organisation which tools aren't working for people.
Is this actually anonymous?Link to this section
Not completely, and it's worth knowing why. We store which answer came from which person. That's how your own answers show in your AI Adoption Portal, and how we make sure you aren't asked the same one-off twice. On an anonymous pulse, nobody in your organisation is shown that link. But in a small group, or from something you write in a comment, someone may still be able to work out an answer is yours.
If it matters to you, check the first screen of each pulse: it says whether your answers are anonymous or identifiable.
If my organisation needs per-person answersLink to this section
It runs an identifiable pulse, which says so before you start and shows your name with your answers. You can close it and answer nothing. On an anonymous pulse, a group where fewer than five people answered shows no figures of its own.
Something missing or out of date here? Tell us.