Show me your hardest prompt problems

If you’re here, you probably work with AI a lot. So here’s a question:

👉 What’s the hardest prompt you deal with regularly?

Examples:

• Super long system prompts that keep breaking

• Multi‑step instructions that different models interpret differently

• Prompts that need to stay consistent across product, marketing, and support​

In Loomi, the goal is to make these “hard prompts”:

• Easier to structure

• Easier to test on Claude / GPT‑4 / Gemini / Llama

• Easier to save and reuse as a clean template​

If you’re open to it, share:

• Your toughest use case

• Which model(s) you’re using

• What “good output” means for you

I can try to:

• Suggest how to structure it

• Share how I’d set it up in Loomi today

• Use your example to improve the product roadmap

Real problems > generic feedback. Let’s make this thread genuinely useful for everyone here.

Try it Now : https://loomiai.vercel.app/