Claude Masterclass
Take Your Skills Further
Go further with Claude: get more helpful answers, read long documents with confidence, build things with Artifacts and Projects, and create routines built around your own work.
What you will be able to do
Included
- 8 modules · 32 short lessons
- 8 module quizzes with explanations
- Downloadable worksheets and workbooks
- Reusable prompt templates you can adapt
- Hands-on projects using documents and Artifacts
- Final assessment and a capstone project
Downloadable worksheets, prompt templates and the capstone kit are unlocked with Pro.
Preview · Module 1 — Getting Better Answers
Free previewLearning objectives
- Decide what reference material a task genuinely needs
- Prepare a document so the answer is grounded rather than generic
- Recognise when missing context is the real cause of a weak answer
The most common cause of a mediocre answer
It is rarely the wording of the request. It is that the model had nothing real to work from, so it produced a competent average of everything it has seen.
With Claude, quality is usually limited by context, not cleverness. If the answer sounds like it could have been written for any company in your industry, the model was missing your material.
- •Your actual document, not a summary of it
- •Your constraints: budget, timeline, audience, risk tolerance
- •Your standards: what you would reject and why
- •Your prior work, so the output matches what you already produce
Before you rewrite the prompt
Ask: what does the model not know about this situation that it cannot reasonably infer? Answer that question first; it is usually faster than rephrasing.
Preparing material the model can actually use
Raw context is not the same as usable context. A fifty-page document with no orientation produces vaguer answers than a well-labelled ten-page extract.
- 1.Label each piece: 'This is the client brief', 'This is last year's report', 'This is our house style sample'.
- 2.Say which parts are authoritative when two documents disagree.
- 3.Remove anything irrelevant — it dilutes attention and adds noise.
- 4.State what you want taken from the material and what you want left alone.
Context-setting pattern
You are helping with [task]. Below are: (1) the source document, (2) our internal standard for this kind of work, (3) an example of output we were happy with. Use (1) as fact, (2) as the rules you must follow, and (3) only for tone and structure. Do not copy content from (3).
Redacting before you share
- •Remove names, addresses, account numbers, and anything health-related unless it is essential and permitted.
- •Replace client identities with labels ('Client A') where the identity is not needed for the task.
- •Check your account's data controls before uploading anything commercially sensitive.
Practical habit
Build a redaction pass into your workflow rather than deciding case by case under time pressure. It takes a minute and removes an entire category of risk.
Exercise
- 1.Pick a task where you have previously been unhappy with the AI output.
- 2.Gather the three pieces of material described above.
- 3.Run the context-setting prompt.
- 4.Compare the answer with the earlier attempt and record what changed.
Key takeaways
- • Weak answers usually mean missing context, not a badly worded request.
- • Label your inputs and say which one is authoritative.
- • Redaction is a routine step, not a judgement call made under pressure.
The full masterclass includes 32 lessons across 8 modules, all four preview lessons above included.
Masterclass curriculum
How to give Claude the right material, ask for the kind of answer you want, and hold the quality steady across a longer session.
Final assessment
End-of-course assessment covering every module, with explanations for each answer.
Capstone project
One real project built across the masterclass, with a brief, checkpoints, and a marking rubric.
Unlock the Complete Masterclass with Pro
Get full access to this Masterclass and everything included with Digital Toolkits Pro.
Cancel anytime · Instant access · Existing Pro members already have this included.