Your prompts are showing

📓 Prompted with phrases like "ketchup in outer space," "ketchup bottle in psychedelic art," or simply "ketchup"
❤️ The look of that iconic bottle came through
😎 What would AI say about your brand?

• YouTube Expands Access to Gen AI Features

• Threads Experiments with AI Interaction Summaries on Profiles

• Facebook will Allow Nicknames When Posting in Groups

The annual State of Business Podcasting Report by One Stone Creative analyzes the top 100 Business podcasts on metrics like show length, release schedule, offers, sponsors, social media, show notes and much more. It’s a snapshot of what the most successful shows are doing, and if you’re considering a podcast, you’ll want to check it out.

Create Prompts with Purpose

We’ve all put a prompt into an LLM and results were what they were.

LLMs have access to far too much information and it’s up to you to give it guidance. You’ve most likely heard the term “prompt engineering,” which simply means the practice of designing and refining prompts to guide an AI model toward generating a desired output. When you create prompts with purpose, your results show.

1. Create personas. You know your target audience. Tell the LLM to build a persona of the person you are trying to reach with your content. What are they reading? What are they searching? What is their job title and expertise? What are their pain points? What solutions can you provide for them? The more specific, the better. Anytime you write a piece of content, tell the LLM to check it against this persona and ensure you aren’t missing the mark.
   

2. Build the right format. Tell the LLM what you want to write about, and ask it for an outline of that idea in the format and length you want. Bullet points, a table, a 500 word blog with section headings, etc. You get to go in and edit to make it your own, but starting in the ideal end-result saves tons of time.
   

3. Prompt in steps. Instead of giving the LLM a huge list of tasks at once, do it step by step. Analyze this research paper, identify key talking points for a speech, recommend an outline to present to an audience of experts.
   

4. Provide examples. Give the LLM already created material that you want it to mirror. If you like the tone of an article, the structure of a blog, the humor in an ad, share that information a better output.
   

5. Don’t use vague phrases. Sentences like “make this sound cool” or “improve this content” isn’t going to net you anything worth it. Tell the LLM exactly how you want something refined.

Precise prompts are good prompts. And as always, AI makes mistakes. Edit, review and look for errors.

Prompts…for Prompts

Use an LLM to help you build a prompt for another LLM.

The only tool necessary to accomplish this is using the knowledge already in your brain.

💹 Give the LLM explicit and detailed context on what you want to accomplish.

 🗒️ Tell the LLM you want it to help you build a prompt in order to get your desired result.

📎 Review the output and if it seems reasonable, have it give you the same prompt but in markdown (headings, lists, code blocks)

🥇 Use the prompt in the LLM that you’re going to work in

And just to stay within our theme, Claude is our favorite tool if you want to give it a try.