daBongo LMS AI Training Courses

Everyday Claude – Core Habits for Productive AI Work

Lesson 5: Building Your Personal Claude Practice

Lesson Objectives

By the end of this lesson, students should be able to:

  • Identify the three highest-value Claude use cases in their specific work
  • Document at least two recurring prompt templates for reuse
  • Apply the "try, learn, refine" feedback loop to improve their practice
  • Know when human judgment overrides AI output – and apply that judgment consistently

Lesson Content

From occasional to deliberate.

Occasional AI use produces occasional value. Deliberate practice – identifying where AI adds the most value, building repeatable workflows, refining prompts over time – compounds. The difference between a user who occasionally gets useful results and a power user is mostly practice and reflection, not inherent technical ability.

Finding your highest-value use cases.

Not all tasks benefit equally from AI assistance. The highest-value applications typically share three characteristics:

  1. High time cost: The task currently takes significant time relative to its output value
  2. Language-intensive: The work involves writing, reading, summarizing, or explaining
  3. Recurring: You do the same type of task repeatedly

Audit your last week. Which tasks were time-intensive, language-heavy, and recurring? Those are your AI candidates. Start there rather than trying to use AI for everything.

Building a prompt library.

For tasks you do repeatedly, save effective prompts. A prompt library can be as simple as a folder of text files or a notes document – one prompt per task type. Each entry contains:

  • Task name
  • The full prompt (including context, task, format, constraints)
  • Notes on what to customize for each use

When you need the task done, open the template, customize the context-specific parts, and paste. This eliminates re-deriving the prompt each time and produces consistent results.

The try-learn-refine loop.

Every time you use Claude for a recurring task, note what worked and what needed editing. Over time, you will notice patterns – certain prompt structures that consistently work, certain omissions that consistently require follow-up. Use this information to update your prompt library. This is the mechanism that makes practice compound: each use teaches you something that makes the next use better.

When human judgment overrides.

No prompt library changes this: for decisions with real stakes, AI output is input – not conclusion. Some examples:

  • Personnel decisions: Claude can help draft a performance review structure, but the judgment of the person's contribution belongs to the manager.
  • Client advice: Claude can draft options, but a professional's judgment on what to recommend belongs to the professional.
  • Creative work with a unique voice: Claude can improve your writing, but your distinctive perspective is yours.
  • Anything with safety or legal implications: Verify with qualified humans.

Building a clear list of where AI is your assistant versus where it is your tool – and where it should not be involved at all – is part of a mature AI practice.

Practical Example

A communications director builds a prompt library over three months.

Starting with a blank document, she adds one entry after each successful Claude interaction.

After ninety days, she has eighteen entries covering: press release drafts, meeting summary formats, stakeholder briefing structures, FAQ drafts, email tones for different situations, and slide content frameworks.

She estimates this library saves her two to three hours per week.

The library is also a training resource: new communications staff receive it as a starter kit rather than starting from scratch.

Lesser-Known Tip

Build a "rejection library" alongside your prompt library – a list of task types where you tried Claude and it consistently produced results that needed so much editing they were not worth the time. These are just as valuable as the successes: they tell you where to stop expecting AI to help and invest that time elsewhere. Not every task should be AI-assisted, and recognizing which ones waste time is a productivity skill.

Safety Notes

A prompt library is a powerful efficiency tool but can become a liability if prompts become stale. When your role changes, when a project ends, or when organizational context shifts, review and update your library. An outdated prompt that provides wrong context can produce confidently wrong output – worse than starting fresh.

Practice Task

Audit the last five days of your work. Identify three tasks that were time-intensive, language-heavy, and recurring. Write a starter prompt for each one. Run each prompt, refine based on the output, and save the refined version as the first entries in your personal prompt library.

Completion Check

You should be able to identify your three highest-value Claude use cases, have at least three prompt templates documented, describe the try-learn-refine loop, and state clearly where in your specific work you will not let AI output substitute for your own professional judgment.

Log in and enroll to access lesson quizzes.

Scroll to Top