Student guide

How to use this workspace to interview stakeholders and uncover what your project really needs.

Last updated 12 September 2026

You are reading this signed out

Everything below applies once you sign in with the access code your instructor gave you. The same guide is inside the workspace under Help. If your code does not work, the sign-in page explains each message you might see.

What this is

This workspace simulates the part of a real systems analysis project where you work out what a system actually needs to do, by talking to the people who will use it. Your team has been given a case: an organisation with a problem, a set of people connected to it, and documents describing how things currently work. Nobody will hand you a requirements list. You have to interview people and work it out.

The people you interview are AI simulations configured by your instructor. They are not real people, and they are not a search engine. Each one knows only what their role would plausibly know. They have opinions, blind spots, and sometimes beliefs that are simply wrong. Part of the exercise is noticing when two people tell you different things.

Treat everything you hear as a claim, not a fact

An interviewee can be incomplete, inconsistent, or mistaken, exactly as in a real interview. Check anything that matters against a second person or a document before you build on it.

Signing in

You need one thing: the access code your instructor gave you. There is no password and no registration step.

  1. Go to the student sign-in page and enter your code in the Access code box.
  2. Codes look like MAPLE-7284: a word, a hyphen, and four digits. Capitalization does not matter: the box capitalizes as you type and ignores characters that cannot appear in a code.
  3. Press Open my workspace.

Your session lasts 30 days, so on your own machine you will usually stay signed in for the length of the project. Use Sign out in the header when you are on a shared computer.

Your code is yours alone

An access code identifies you personally, not just your team. Every question you send is recorded under your name, so sharing a code means someone else’s work is attributed to you, and yours to them.

If the code is rejected

  • “That access code is not valid” means the code was mistyped or is not recognised. Check it character by character.
  • “This enrollment is no longer active” or “This class is not currently active” means your instructor has to reactivate you or the class.
  • After several failed attempts in a short time the page asks you to wait a few minutes before trying again.

Finding your way around

The header always shows your case, your class, your team and your name. Four links sit beside it.

The four links in the workspace header
WorkspaceYour home base: who you can interview, the case briefing, your team, and instructor updates.
DocumentsEvery document your team currently has access to. New ones appear here as you uncover them.
Voice settingsYour personal speech preferences. These affect only you, not your teammates.
HelpThis page.

A Stakeholders column lists everyone your team knows about. People you can interview are clickable. People marked Not yet available are known to exist but cannot be interviewed yet.

If your instructor has paused interviewing, a banner says so. Transcripts and documents stay readable while it is paused.

Interviewing someone

  1. Click a person on the workspace page or in the stakeholder column.
  2. Read their role and background at the top. It tells you what they are likely to know about.
  3. Type your question in the box and press Send. Enter sends; Shift and Enter together add a new line.
  4. Their reply appears in the transcript, and is added to the same transcript your teammates see.

Questions are limited to 4,000 characters, which is far more than you should need. A counter under the box shows how much you have used. Ask one thing at a time: a question with five parts usually gets one part answered well and four answered vaguely.

Use the search box above the transcript to find something said earlier. Refresh pulls in anything your teammates have asked in the last few moments.

Everything you send is permanent and shared

There is no way to delete or edit a message once it is sent. Every teammate sees your questions with your first name attached, and your instructor can read the full transcript.

Waiting, queues and errors

Only one question per person is answered at a time, for the whole team. If a teammate asked something a second before you, yours waits its turn. A small coloured label next to the person’s name tells you what is happening.

What each status next to a stakeholder's name means
QueuedWaiting behind another question. It may show your position in line.
Thinking…Your question is being answered right now.
Listening…The microphone is on and capturing your speech.
Transcribing…Turning what you said into text.
Speaking…Reading a reply aloud. This never changes the text.
ReadyNothing in progress.
ErrorSomething went wrong and nothing was saved. Your question is still in the box, and the message says what happened and how long to wait — a minute if the stakeholders are simply busy, longer if the service is not responding. If an error keeps coming back, tell your instructor.

Keep the tab open

Questions move through the queue while the page is open. Closing the tab immediately after sending can leave your question waiting. Give it a few seconds.

When something goes wrong

A failed question is never added to the transcript, and your typing is kept in the box so you can simply send it again. Two messages are worth recognising.

  • Still waiting. The answer is taking unusually long. It may yet arrive, so press Refresh before retyping.
  • Your question was not answered. Something failed. Nothing was recorded. Send it again.

There are limits on how fast questions can be sent: roughly ten a minute for you, and a few hundred an hour for your team. These exist to stop something going haywire and sending thousands of questions, not to ration your thinking. Normal interviewing will never come close.

Finding people and documents

You will not be shown everyone at the start. Additional people become available when your team asks about the right topic with someone who would plausibly refer you onward. This mirrors real fieldwork, where one interview earns you the next.

Documents work the same way. Some are given to you at the start, some are released by your instructor, and some are handed over by a person during an interview when you ask about the subject the document covers. When something new arrives, a short notice appears in the corner and the sidebar updates.

If you feel stuck, you probably have not asked the right person

The workspace hints that not everyone has introduced themselves. When a person mentions a colleague, a team, a report, or a system you have not heard of, that is a thread worth pulling. Ask them about it while you have them.

People you have unlocked normally stay available, though your instructor can change that if the case needs it. Documents can be withdrawn from view too, so open and read what you are given rather than assuming it will be there later.

Messages from your instructor

Your instructor can post updates to the whole class. Important ones appear at the top of your workspace and every interview page until you press Got it. Acknowledging is per person: clearing it for yourself does not clear it for your teammates.

Some updates are marked Correction. These override what a stakeholder previously told you. If a correction contradicts something in your transcript, the correction is what counts.

Every update, read or unread, stays listed under Updates from your instructor at the bottom of the workspace.

Working as a team

You share one transcript per person with your whole team. Everyone sees every question and every answer, labelled with the asker’s first name. Nobody has a private conversation.

  • Read what your teammates have already asked before starting. The sidebar shows a message count for each person.
  • Divide people between you rather than duplicating interviews. Repeated questions make everyone wait longer and rarely produce new information.
  • New messages from teammates appear automatically within about twenty seconds, or immediately with Refresh.
  • Because the transcript is shared, a good question benefits everyone. A vague one clutters the record for everyone too.

Your team can name itself. Use Rename on the team card of the workspace page; the name must be unique in your class and is visible to your teammates and your instructor, who can change it.

Speaking and listening

Speech is entirely optional. Every reply is always shown as text, whether or not it is read aloud, and recognised speech always lands in the message box for you to check before it is sent.

Hearing replies

Read stakeholder responses aloud is on by default. Speech rate adjusts how fast, from half speed to double. Stop speaking in the interview header halts playback without affecting the transcript.

Asking by voice

Choose how the microphone button behaves under Voice settings.

The two microphone modes
Click to start, click to stopThe default. Press once to begin, once more to finish. Works from the keyboard with Space or Enter, and is the better choice for accessibility.
Press and hold to talkHold the button down while speaking and release when finished, like a walkie-talkie.

Automatically send when I finish speaking is off deliberately, so you can correct a misheard word before sending. Turn it on only once you trust the recognition.

If your browser blocks the microphone, allow it in your browser’s site settings for this page. If voice input is unavailable altogether, the app says so and you can type as normal.

Installing better voices

There are two kinds of voice in the settings list. Natural voices are generated by this app on your own computer and sound markedly better than anything most machines ship with. Everything else comes from your operating system, so what you get depends on your machine.

Start here: pick a natural voice

Open Voice settings and choose one from Natural voices, then press Test voice. The first time, your browser downloads the voice once (92 MB) and keeps it. After that it loads in about a second, a reply starts speaking within a few seconds, and it works with no internet connection at all. Every student hears the same voice whatever computer they use, and the speech is produced on your own machine, so nothing you hear is sent anywhere.

Do the download once, on a decent connection, before you need it. On a slower or older computer the voice takes longer to start speaking; if that bothers you, a device voice will always be quicker.

The rest of this page is for adding better device voices, which is worth doing if you would rather not download anything, or if the natural voices are slow on an older computer.

Speech is optional either way

Every stakeholder reply is always shown as text. Reading aloud is a convenience, so skip all of this if you prefer to read.

macOS

Apple calls the good voices Enhanced and Premium. Premium is the best. Each one is a few hundred megabytes, so use a decent connection.

  1. Open System Settings, then Accessibility.
  2. Choose Read & Speak. On macOS 15 and earlier this same section is called Spoken Content instead. If you cannot spot it, type Spoken Content into the search box at the top of System Settings, which finds it under either name.
  3. Click the small information button (ⓘ) next to System voice, then Manage Voices…
  4. Expand English and find a name with (Enhanced) or (Premium) after it. Click the download arrow beside it. Good choices are Ava, Zoe, Allison, Evan, Nathan and Serena.
  5. Wait for the download to finish, then quit your browser completely and reopen it.
  6. Return to Voice settings in this app and pick the new voice.

Apple renamed this section in macOS 26 (Tahoe). Both names are given above because the instructions have to work on whichever version you are running.

Windows

On Windows the browser matters more than anything you install. Microsoft’s best voices are wired into Edge and are not shared with other browsers.

Easiest fix: open this app in Microsoft Edge

Edge on Windows 11 offers Microsoft’s natural voices with nothing to install. They are the best voices available on Windows by a wide margin. They are generated online, so they need an internet connection to speak.

If you would rather stay in Chrome or Firefox, those browsers can only see the older Windows voices, and you can add more of them:

  1. Open Settings, then Time & language, then Speech.
  2. Under Manage voices, choose Add voices and install the English voices you want.
  3. Restart your browser, then pick the new voice in this app.

Do not bother with the Narrator route

Windows also offers natural voices under Settings, Accessibility, Narrator. Those are reserved for Narrator and a few Microsoft apps: Microsoft does not expose them to other programs, so they will not appear in Chrome or Firefox however many you install. Use Edge instead if you want them.

ChromeOS and Chromebooks

  1. Open Settings and search for Text-to-Speech. It sits under Accessibility.
  2. Open the text-to-speech voice settings and install any additional English voices offered there.
  3. Reload this app afterwards.

Chromebooks have the fewest options, and the exact wording of these settings changes between ChromeOS releases, so search rather than following a fixed path. Google’s network voices, which Chrome provides on any platform, are usually the best available on a Chromebook and need an internet connection in order to speak.

Choosing a voice in the app

The list is filtered to English voices and grouped by how clear they are. Voices under Clearest on this device are worth trying first. Two kinds are hidden: joke voices such as Bad News, Bubbles and Zarvox, which are sound effects rather than speech, and the oldest robotic voices such as Fred and Albert, which macOS keeps only for compatibility.

  • Press Test voice to hear a voice before committing to it.
  • Adjust Speech rate if a voice is clear but too fast or too slow.
  • Automatic (best match) picks the closest voice to each stakeholder’s configured language and style, so different people can sound different. Choosing a specific voice makes everyone sound the same.
  • Your choice is saved to your account rather than to the browser. Voice names differ between devices, so if you sign in elsewhere and that voice is missing, the app says so and falls back to the automatic match.
  • If you have already chosen a voice, it stays in your list even when it is one the filters would normally hide.

If nothing changes

  • Quit the browser entirely and reopen it. Browsers read the voice list once at startup.
  • Confirm the download actually completed. Large voices can fail quietly on a slow connection.
  • Check your operating system can use it: on macOS, the voice should now be selectable under System voice.
  • Try a different browser. Voice availability differs between Chrome, Edge, Safari and Firefox on the same machine.

Saving a copy of your work

The workspace page offers Download Markdown and Download CSV. Both contain every conversation your team has had with every person, in order.

  • Markdown is for reading, quoting in a report, or pasting into a document.
  • CSV is for analysis in a spreadsheet: coding themes, counting who said what, tracing where a requirement came from.

Export before you write anything up. Being able to cite the exact words a stakeholder used is far more convincing than paraphrasing from memory.

Interviewing well

The simulation rewards the same habits a real interview does. A few that make an immediate difference:

  1. Open before closed. “Walk me through how an order gets processed today” yields more than “Do you use a spreadsheet?”
  2. Ask for the last time it happened. Concrete recent examples surface details that general questions never reach.
  3. Follow the pain. When someone complains, ask what it costs them, how often, and what they do instead. That is where requirements live.
  4. Chase every proper noun. A named person, system, form or report is a thread worth pulling.
  5. Ask the same question of two people. Disagreement between roles is a finding, not a problem.
  6. Check what you think you heard. Summarise it back and ask whether you have it right. People correct a wrong summary more readily than they volunteer a missing detail.
  7. Ask what you have not asked. Closing with “What should I have asked you that I did not?” is reliably productive.

Interviewees answer from their own role and perspective. Asking a warehouse supervisor about board-level strategy will get you a warehouse supervisor’s view of it, which may be exactly what you need, or may mean you are asking the wrong person.

Troubleshooting

Common problems and what to do about them
Nobody is listed to interviewYour team may not have a case assigned yet, or the first person is still to be released. Read the briefing and check Documents.
A person is greyed outThey are known but not yet available. Keep interviewing the people you can reach; they lead to the others.
My question vanishedFailed questions are never recorded, and your text stays in the box. Send it again.
It has been stuck on QueuedA teammate's question is ahead of yours. If it lasts more than a minute or two, press Refresh.
Interviewing is disabledYour instructor paused it. Transcripts and documents still work.
A document link does not openIf it reads “Link unavailable”, no address has been set yet. Tell your instructor.
Every voice sounds roboticYour device only has its basic voices. See Installing better voices above.
My chosen voice disappearedVoice availability differs between browsers and devices. The app falls back to the closest automatic match and tells you it has done so.
The microphone does nothingCheck the browser has permission for this site, and that no other application is holding the microphone.

If a stakeholder breaks character, contradicts itself inside a single answer, or describes something that clearly does not belong in this case, tell your instructor. Someone being guarded, vague, or unwilling to speculate is not a fault: people are like that, and getting past it is the skill being taught.