Improving Personalization for Event App Creation: Seeking Advice

Theo Kanellopoulos · 2025-01-20 17:02:37.289998+00

hi team, long time without a question :slightly_smiling_face:, I’m trying to create a personalization and some times it works really well and some other times it doesn’t work so well. what would you advise for me to do to improve the result? I tried using your prompt generator but the outcome was worse than my prompt. my example of a good message for a university for example would be: “create an event app to help students get info on their first day in the school”, something that feels that I have done deep research.

Improving Personalization for Event App Creation: Seeking Advice | Clay

5 comments

· Sorted by Oldest

Channeled
APP
·
·
Here are key steps to improve your personalization results: 1. Start with a clear system prompt that provides context and outlines your specific goal for the university event app. 2. Follow these essential practices for better results: • Provide 3-5 examples of the style you're aiming for • Include what you don't want (negative exclusion criteria) • Specify formatting details and desired tone • Give clear instructions on how to structure the content 3. Make your instructions step-by-step and explicit to prevent the AI from making incorrect assumptions. 4. Be specific about how you want the output formatted and returned, including details about tone, word count, and style.For your university event app example, try expanding your prompt with more specific details about the first-day experience, target audience needs, and desired information structure. This will help the AI generate more focused and relevant content. Sources: Prompt Engineering Crash Course [Limitless Research] - Clay University Prompt Engineering Crash Course [Clay 101] - Clay University
Channeled
APP
·
·
Hi Theo Kanellopoulos! This thread was recently closed by our Support team. If you have a moment, please share your feedback:
Channeled
APP
·
·
Thanks! We've reopened this thread. You can continue to add more detail directly in this thread.
Arturo (Clay)
·
·
Hey Theo, thanks for reaching out!! Long time no see! Will give a few of these a shot. Of course, I can share that I've seen/heard people recommend Claude models for message generation as it tends to output more 'natural' drafts if you'd like to compare them with the same prompt, for example. Any columns in particular you're trying to improve though?
Channeled
APP
·
·
Hi Theo Kanellopoulos! This thread was recently closed by our Support team. If you have a moment, please share your feedback:

Channeled
APP
·
·
Here are key steps to improve your personalization results: 1. Start with a clear system prompt that provides context and outlines your specific goal for the university event app. 2. Follow these essential practices for better results: • Provide 3-5 examples of the style you're aiming for • Include what you don't want (negative exclusion criteria) • Specify formatting details and desired tone • Give clear instructions on how to structure the content 3. Make your instructions step-by-step and explicit to prevent the AI from making incorrect assumptions. 4. Be specific about how you want the output formatted and returned, including details about tone, word count, and style.For your university event app example, try expanding your prompt with more specific details about the first-day experience, target audience needs, and desired information structure. This will help the AI generate more focused and relevant content. Sources: Prompt Engineering Crash Course [Limitless Research] - Clay University Prompt Engineering Crash Course [Clay 101] - Clay University
Channeled
APP
·
·
Hi Theo Kanellopoulos! This thread was recently closed by our Support team. If you have a moment, please share your feedback:
Channeled
APP
·
·
Thanks! We've reopened this thread. You can continue to add more detail directly in this thread.
Arturo (Clay)
·
·
Hey Theo, thanks for reaching out!! Long time no see! Will give a few of these a shot. Of course, I can share that I've seen/heard people recommend Claude models for message generation as it tends to output more 'natural' drafts if you'd like to compare them with the same prompt, for example. Any columns in particular you're trying to improve though?
Channeled
APP
·
·
Hi Theo Kanellopoulos! This thread was recently closed by our Support team. If you have a moment, please share your feedback: