> For the complete documentation index, see [llms.txt](https://breakpoint-journal.gitbook.io/breakpoint/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://breakpoint-journal.gitbook.io/breakpoint/ai/claude-and-ai-fluency.md).

# Claude and AI Fluency

## The 4D framework for AI fluency

One useful model for working well with AI is the 4D Framework for AI Fluency. It comes from a research collaboration between Professor Rick Dakan at Ringling College of Art and Design and Professor Joseph Feller at University College Cork.

The framework focuses on four core skills:

* **Delegation**: Decide what you should handle, what AI should handle, and how to divide the work. Match each task to your goal and to the model's strengths.
* **Description**: Communicate clearly with the AI. Define the output, guide the process, and state the behavior you want.
* **Discernment**: Review the result critically. Check accuracy, quality, fit, and what still needs revision.
* **Diligence**: Use AI responsibly and ethically. Be transparent, make careful choices, and stay accountable for AI-assisted work.

Together, these four skills help you get more value from every AI interaction.

### 4D exercise

> **Communication project**
>
> You're working with an AI assistant to draft a series of emails for a marketing campaign.
>
> **Delegation**: What parts of this project would you own yourself, and what would you do with AI?
>
> **Description**: How would you explain the campaign's tone, goal, and success criteria to the AI?
>
> **Discernment**: What standards would you use to judge whether the AI-written emails meet your needs?
>
> **Diligence**: What should you consider around transparency, responsibility, and oversight?

## Tips for using Research well

The **Research** feature works best when you treat the prompt like a brief. Since runs can take anywhere from 5 to 45 minutes, a stronger prompt usually leads to a better result.

* **State the goal clearly**: Avoid broad asks. Instead of asking for a general overview, define the exact decision, market, or question you want analyzed.
* **Give Claude a structure**: If you want the output split into sections, say so up front. Research will organize findings around the outline you provide.
* **Add constraints early**: Include budget, timeline, geography, audience, or other limits. These details help narrow the search and improve relevance.
* **Refine the prompt first**: If the research question still feels fuzzy, ask Claude to help you draft a better Research prompt before you start the run.

In short, the more direction you provide, the more useful the final research will be.

## Resources

If you're interested in discovering more feel free to check out these short courses:

* Anthropic's [AI Fluency](https://anthropic.skilljar.com/ai-fluency-framework-foundations)
* Anthropic's [Claude 101](https://anthropic.skilljar.com/claude-101)