🧠 LLM Use Case Guide
Simon Wang | 🌐 Checkout my Website!
Want to use AI for a task but not sure where to start? This quick guide helps you choose the simplest approach for your needs. It's best practice to start with the easiest method that works, then explore more advanced options if needed to avoid wasted effort!
These are general recommendations and aren't absolute guidelines to every use case! This guide is meant as a good starting point to figure out what tool to try, but you might find your specific use case or preferences to be a bit different.
And that's perfect! Using AI is all about creativity and ingenuity, the more deeply you understand your use case and requirements, the better you can leverage this powerful tool in unique ways! 🤩
🤔 What do you want the AI to do?
These are just some common use cases to give a good basic guide for what different LLM-based approaches are good for. If you have a more unique use case, try out the custom input feature below that uses an LLM to give you suggestions!
💬 Describe Your Use Case
This feature relies on using an LLM (gpt-4o-mini) to give recommendations based on a crafted system prompt, but it is NOT perfect! It is still a good starting point, and the better and more in detail you can describe your use case the better the suggestion will be!
✨ When to Try More Advanced Tools
Here are situations where advanced LLM techniques might be worth exploring:
Use when you need the AI to answer questions based on content it doesn’t know, like PDFs or company docs.
Use when your task has multiple steps or needs planning, memory, decision-making, or access to external tools.
Use when the base model isn’t good enough for your domain, tone, or style. When you need the AI to more reliably follow your specific requirements!
⚖️ Complexity vs. Control
Simple prompts are quick and flexible. But as tasks grow more specific or complex, more advanced methods offer more control — at the cost of time and setup.
- 🟢 Prompting → minimal setup, fast results
- 🔵 RAG → better factual accuracy with your data
- 🟡 Agents → complex workflows, reasoning, autonomy
- 🔴 Fine-Tuning → highest control, requires data + compute