3. Getting Better Results

Course: Claude 101 Estimated time: 13 minutes

Learning objectives

By the end of this lesson, you will be able to:

Troubleshoot common challenges when working with Claude
Understand what AI fluency means and how it improves your results
Evaluate whether Claude is a good fit for a specific workflow or task
Apply techniques that make your Claude experience smarter over time

Common challenges and how to fix them

Challenge	Fix
Too generic	Add more details about audience, role, or constraints
Too long	Ask to "summarize in 3 bullet points" or "keep under 200 words"
Lacks nuance	Provide examples of what good looks like
Wrong tone	Specify the tone explicitly ("formal," "conversational," "direct")

The frustration mindset

One of the most significant parts of effective Claude use is recognizing a response you don't like — then finding a better approach.

First drafts are great starting points. Don't abandon Claude after one try; refine
Going deeper — when Claude doesn't go deep enough in the first response, ask it to go deeper, add more examples, or consider more edge cases
More context — it can help Claude if you provide additional context. Sharing relevant background
Know when to start over — Sometimes, a new conversation is better than a convoluted one

What is AI Fluency?

AI fluency is the ability to collaborate effectively with Claude — mindset-oriented thinking related to using AI tools, but wrapping the knowledge mindset in four key skills:

Delegation — knowing when and what to ask Claude for help with, identifying what kinds of tasks AI is good at vs. what it's bad at, and knowing your own strengths vs. Claude's strengths
Description — giving Claude enough context to do what you want, in the format and style you want
Discernment — critically evaluating Claude's outputs, knowing when it's good enough vs. needing editing
Diligence — following up, iterating, and staying involved in the outputs (versus simply accepting the first thing Claude gives)

Evaluating Claude for your workflows

To evaluate whether Claude would be useful for a specific task:

Trial and error — bring 5 to 10 examples of real work you already do and run them through Claude
Find a match — ask yourself: "does Claude's output match the quality I'd need?"
Add iteration — add one round of follow-up to your test to see if Claude can improve

A simple experiment:

Choose one task you complete regularly (write emails, draft documents, compile research)
Give Claude that task with your normal context
Evaluate: could this replace a first draft? Could this save you time on the research stage?

A smarter experience over time

Claude gets smarter as it gets to know you. There are a few ways to build that context:

Provide context in your system prompt or at the start of conversations
Use Projects for ongoing work — Claude retains context across conversations in a Project
Add instructions in your Project settings for persistent behavior

Lesson reflection

Before moving on, consider:

What types of prompts frequently produce results you don't love?
Which of the four AI Fluency competencies (Delegation, Description, Discernment, Diligence) do you feel least confident in?
What's one task you could trial with Claude this week?

Getting better results