I’d Like Some More, Please
When I first added AI to my workflow I was using ChatGPT exclusively. Then I began experimenting with Claude. Here’s what I’ve learned after thousands of prompts: each AI model has different strengths. Now I have an AI team. My workflow incorporates multiple models for the same task. I might start by drafting my content or code snippet in ChatGPT. I’ll take that outcome, add my changes and enhancements and get a clean version from ChatGPT (grammatical errors? So 2023!) Then it’s off to Claude for additional refinements to the tone, flow, or to catch logic gaps. And depending on what I’m doing - writing, developing code, gathering information - I might take it to Manus or Gemini. Copilot gets a nod too if you’re developing code. It’s wicked fast. The power is not in one AI model, it’s taking the best from using them all. An example, when writing code you would never ship a product after just one person looked at it. Same principle applies here. Each model brings a different perspective. Here’s my simple framework: 1. Generate in your primary model to get the “first draft”. 2. Refine the flow and tone in a second model to catch what you and the first model missed. 3. If necessary, or you are just curious (and I am most of the time) take a third pass. Especially if something specific needs refinement (creativity, technical accuracy, etc) Also, never (never) just copy and paste. This is YOUR work. Be present in the finished product. Your expertise is what ties it all together. Final result? Way stronger than any single model could have produced. And I gain a lot more in skills and even perspectives. TLDR: If you’re only using one AI, you’re leaving value on the table. These tools are free or cheap. A multi-model workflow is a competitive advantage. How do you optimize the power of working with AI? What are your thoughts on using multiple AI models? All the best.