Questions
Your questions, answered.
How a scenario runs, how your score is worked out, where your data goes and what happens if you step away part-way through. Plain answers, including what the AI does and doesn’t decide.
01
Getting started
What a scenario is, who it’s for and how much time it takes.
What is RoleCube?
RoleCube is a place to practise a job before you have it. You step into a multi-day workplace scenario and work through stakeholder messages, meetings, a data workspace and documents. Then you write and submit a recommendation. You get back a manager-style evaluation of the decisions you made and the evidence you used.
The companies, people and datasets are fictional. The uncertainty is real: the evidence is incomplete on purpose, and there’s no hidden correct answer to work backwards from.
Who is RoleCube for?
Anyone who wants to practise the judgement a role needs, not just read about it. That includes people preparing for a first role, moving into a new one, or early in a job and wanting somewhere a wrong call is cheap. Scenarios are currently written for Data Analyst and Product Manager roles, each at a stated seniority, and each one lists the skills it covers before you start. See Roles for what each one covers.
Which scenarios can I play right now?
There are ten scenarios for each of two roles today: Data Analyst and Product Manager. They range from easy to expert and from junior to director level. Each one runs over three simulated working days and has its own company, cast, documents and data. Roles shows what each role covers, with an example scenario and its day plan, so you can see what you’re signing up for before you create an account.
How long does a RoleCube scenario take?
Each scenario runs across several simulated working days (the ones available today are three days long), and each day has a clock from 08:00 to 18:00. How long that takes in real time is mostly up to you. In Profile you can run the clock at 1× up to 12× simulated hours per real hour, or switch the timer off, and you can end a day whenever you think you have done enough.
Most of the time goes on reading, querying and deciding, so it varies from person to person. You don’t have to finish in one sitting. See what happens if you stop part-way.
Do I need to know SQL or have done the job before?
You don’t need to have done the job. The data workspace takes SQL-style queries against the scenario’s own read-only tables, and scenarios suggest queries to start from, so basic familiarity helps a lot. If you get stuck, progressive hints are there when you ask for them. Each scenario is pitched at the role and level it states, so it should stretch you without assuming you already know the answer.
Is RoleCube free?
Yes, for now. There’s no payment step and you can’t be charged. Every account is on a plan, given by the access code you sign up with while RoleCube is invite-only. Some plans, such as a beta or promotional one, can limit how many scenarios or attempts you can start, and when a plan ends you can still sign in and see your results, but you can’t start new scenarios until you’re on a plan again. If we ever introduce paid plans, we’ll publish their prices and terms first. See section 5 of the Terms and Conditions.
02
How grading works
A rubric written with the scenario, AI help with the judgement, and arithmetic done in code.
How is my RoleCube submission graded?
In two layers, combined into one score:
- Process, computed in code from what you did: stakeholders briefed, meetings attended, documents opened and evidence attached. Hints, overtime and lateness appear in your feedback but never reduce your score.
- Rubric: each criterion written with the scenario is scored 0–4 against its answer key, with AI assistance.
The two are blended using weights set for that scenario. The blend runs in code, so the final number can always be traced back to its parts. The full breakdown is on the Product page.
What does the AI do, and what doesn’t it do?
RoleCube uses Google’s Gemini in a few specific places. It plays the stakeholders who reply to your messages and answers your follow-up questions after a meeting, both based on what was written for that scenario. It scores each rubric criterion against the answer key, reading your whole submission. And it writes the plain-language manager review.
It doesn’t decide the facts of the scenario, what a stakeholder knows, the results of your queries or the arithmetic of your score. It’s given the scenario’s own material and told to restate or judge only against that.
If I pick the right root cause, do I get full marks for it?
Not automatically. You choose the root cause from a list, so a correct pick on its own doesn’t show your reasoning. Your own writing has to support the cause you chose. Where it doesn’t, or where you say you guessed, credit above a small participation floor is reduced in proportion. A correct pick is never marked lower than partly met.
What happens if the AI is unavailable when I submit?
Your submission and evaluation still go through. Every AI path has an authored fallback, so messaging, meeting questions, submission and evaluation keep working without it. The evaluation falls back to the deterministic score, is flagged for a human to review and has no AI-based adjustment. An outage is never meant to cost you points.
What happens straight after I submit?
Your report is saved immediately and the scenario is complete at that point. Nothing else is needed from you. The evaluation is generated just afterwards, so it can take a short while to appear. The results page tells you it’s being written rather than leaving you on a blank screen, and refreshing brings it up once it’s ready.
From then on it lives in your results, at a permanent address for that attempt, with the score, the rubric breakdown, the outcomes you met and the written review.
Do hints count against me?
Hints are recorded and appear in your evaluation, so you can see how independently you reached your conclusion. The manager review is told to mention them constructively, not to penalise you for asking.
Will my score change if I open the evaluation again?
No. The evaluation is saved with the attempt, including the rubric it was graded against, so it reads the same every time and is never re-graded. Replaying a scenario starts a fresh attempt and keeps the earlier evaluation in your results.
03
Data & privacy
The short version, and where to find the full details.
What personal data does RoleCube collect, and why?
The Privacy Notice has the full answer: what we collect, why and on what legal basis, how long we keep it, and who operates RoleCube. We link to it here instead of summarising it, so there’s only one version to rely on.
Is what I write sent to an AI provider?
Yes. The messages you send, the questions you ask in meetings and the report you submit are processed by the AI provider to write stakeholder replies and grade your work. Your name, email and profile are never sent. While RoleCube uses Google’s free Gemini service, Google may use what you type to improve its products and its staff may read it, so never include real personal or confidential information. Which provider receives what, and under what terms, is set out in the Privacy Notice and the sub-processor list.
What cookies does RoleCube use, and can I change my choice?
Every cookie and browser-storage key, what it does and whether it needs your consent is listed in the Cookie Policy, which also explains how to change a choice you have already made.
How do I access, correct or delete my data?
Your rights, and how to use them, are in the Privacy Notice. It’s the page to read first, and it explains how to contact us with a request.
04
Account & technical
Progress, unfinished sign-ups and what you need to run it.
What happens to my progress if I stop part-way through a scenario?
It’s kept. Your day, clock, messages, meetings, opened documents, query history and report draft are saved to your account as you go (the report draft autosaves while you type), so you can close the tab and pick up the same attempt later on any device.
Coming back to the scenario opens your latest attempt where you left it. You can’t restart an attempt part-way through. That’s deliberate, so every review is a fair read of how you handled the whole week. Once you’ve submitted, you can run the scenario again from Scenarios, which starts a fresh attempt and keeps the earlier one in your results.
My session expired, or I followed a link while signed out. Where do I land?
Where you were going. Sign-in remembers the page you asked for, including the exact section, and takes you back there once you’re signed in, after a quick check that your profile is complete.
I started signing up but didn’t finish. Do I have to begin again?
No. If your account was created but your profile was never completed, signing in takes you back to the step where you stopped, skipping account creation, and then on to where you were heading. The answers you had already given are saved to your account as you go, so this works on any device.
If you never verified your email, signing in takes you straight to verification and sends a fresh code.
Do I need a code to sign up?
Yes. RoleCube is invite-only for now, so you enter an access code on the first page of sign-up, before an account is created. If you follow the link you were sent, it fills the field in for you. The code puts your account on the plan it names.
A code that’s wrong, expired or already used up is rejected with a message, and you can’t continue until it’s corrected. Codes can currently be entered only when you sign up, and each is meant for one person, so please don’t share yours. See the Help Centre if a code won’t work.
Why does RoleCube say I have reached my plan’s limit?
Your plan limits how many scenarios or attempts you can start, either in total or within a repeating period such as a week or a month. The message says which limit you have reached and, when it resets, the date. An attempt counts towards a limit once it has started, whether or not you finished it.
A limit never stops you continuing an attempt you have already started, and every result you already have stays available, including after a plan ends. With no active plan you can’t start new scenarios.
What do I need to run RoleCube?
A current version of Chrome, Edge, Firefox or Safari, with JavaScript and cookies turned on. Blocking the session cookie prevents sign-in. It works on a phone, but the data workspace and report editor are built for a larger screen, so a laptop or desktop is strongly recommended. Audio is used in meetings, and a transcript is always available. The Help Centre covers troubleshooting.
Ready to give it a go?
Still have a question? The Help Centre goes deeper on using the product.