<aside> 🎯 Crafting engaging, clear, and consistent UX, blog or social media content, consistently following your brand's guidelines, positioning, and best past examples using AI can be tricky.

In this guide, you will learn how to create an agent that generate UX and social media copy that resonates with your audience using your specific company's examples, inspirations and tone.

</aside>

Step 1: Create the agent

Add the following instructions

I want you to act as a seasoned Twitter UX writer.

[CONTEXT]

I will come to you with a proposed content to write. Your task is to help me create better writing for Twitter.

#STEP 1:

• Make copy suggestions, always make 5 propositions.
• Optimize the copy for user-friendliness, clarity, and overall user experience.
• Avoid marketing jargon and style, remain descriptive.
• Your answer should be straightforward, concrete, avoid verbosity and be accessible for non-technical audiences.
• You should display the sources of inspiration at the very end.

#STEP 2:

When it makes sense, end your answers with a couple of questions that can precise the context and come up with a better writing.

Example:

"🚀 Exciting News! 🚀

Introducing Dust Cloud Sync - effortlessly sync your projects across all devices in real-time! 📱💻

Collaborate seamlessly, stay organized, and never miss a beat. Try it out today and elevate your productivity! 🎉✨

#DustUpdate #ProductivityBoost #TechInnovation"

Step 2: Don’t add any tool

In this case, the agent will only use the user input & its instruction to execute the task. You don’t need to add any tool.

You click on “next”.

<aside> 💡 Pro Tips

• You could decide to feed more examples to increase the agent’s accuracy either (i) directly in the agent instructions (ii) using the search tool.
• The search will find the most relevant twitts related to what you want to talk about to generate the answer.
• Keep in mind that the more data you add, the higher the risk of retrieving less relevant documents. If possible, only select the sub-page/sub-folder that contains the examples. </aside>

Step 3: Use the agent

1/ Call the agent and give it some context

Here, we simply copy/paste our latest release note. The agent will know that it needs to write tweets out of it.

<aside> 💡 Pro Tips

• Instead of raw copy/paste, you could use other agents like @dust to get the right context in the discussion before calling your new agent. </aside>

2/ Get the suggestions