Precept One shop mystery. A new way to think.

AI can give you an answer. Can you tell if it’s right?

Precept is one interactive story built around a money mystery. Marnie's sales report shows £1,440.59 in revenue, but the trusted total is only £800.69 — somewhere, £639.90 is being counted incorrectly.

Follow the clues, find the mistake and help Marnie decide what to do next. Along the way, you will learn how to ask better questions, check AI’s work and make sense of numbers. No prior experience is needed.

Check the numbers Follow the clues Save the repair Make it repeatable Explain the decision

Five acts, thirty missions, one case

  • Act 1 — Spreadsheet Sense

    Formulas

    Help Marnie lock references, count filled cells correctly, repair broken SUM ranges, inspect an outlier, write a useful IF, and choose an honest chart.

  • Act 2 — From Cells to Queries

    Structured Query Language (SQL)

    SQL is a language used to ask questions of information stored in a database. Turn familiar spreadsheet questions into clues that explain the missing money.

  • Act 3 — Time Machine

    Git

    Git records changes to files so work can be reviewed, restored and combined safely. Use it to protect Marnie’s repair without losing earlier work.

  • Act 4 — Reliable Workshop

    dbt (data build tool)

    dbt uses SQL to clean, organise and check data in a repeatable way. Use it to stop the £639.90 mistake from returning.

  • Act 5 — Crystal Ball

    Business Intelligence (BI)

    BI uses data, charts and dashboards to help people understand what is happening and make decisions. Turn the repaired numbers into useful advice for Marnie.

What you will be able to do

  • Ask useful questions of stored information
  • Fix broken spreadsheet calculations
  • Save and combine changes without losing work
  • Make a trusted calculation repeatable
  • Explain why a revenue figure can be trusted

Common questions

Do I need any prior experience?

No. The course starts from familiar spreadsheet ideas, maps them into SQL, and builds up to Git and dbt. It is designed for complete beginners.

What tools will I use?

You begin with a familiar spreadsheet. Each later tool is introduced in plain language when the story needs it, with hands-on practice in your browser.

How long does it take?

Three to four hours at a fully guided pace, or 60 to 90 minutes if you are refreshing skills you already know. Each mission takes 5 to 12 minutes and is easy to pause.

Why is the case about revenue?

Every skill is taught because Marnie's case needs it. You follow one £639.90 discrepancy from the first clue to a decision supported by clear charts and evidence.