Liza Adams

How do you know which ChatGPT model to use? GPT-4o. GPT-4.5. o3-pro. o4-mini-high.

Unless you’re following every update, it’s hard to know what to pick or why it matters.

Yesterday I shared how techie language can hold people back from using AI. This is part two of the same problem: even the choices feel unclear.

A few months ago, I built a custom GPT (AI we can build to do a specific task) to help with that. Since then, even more models have rolled out. So I made a new one.

I asked ChatGPT to help me build a GPT inside ChatGPT that recommends which ChatGPT model to use.

It’s a little ridiculous but also the easiest way to get it done.

It’s called the AI Model Advisor for GTM.

It’s built for marketers, sales reps, ops, and CS teams who want to:

• ➡︎ write better messaging

• ➡︎ plan smarter

• ➡︎ move faster on daily work

• ➡︎ build or automate something

…without guessing which model is right.

You answer a few simple questions. It recommends a model in plain English.

You also get a cheat sheet that shows the differences with examples.

This is one small step to make ChatGPT easier for more people to use and explore.

I hope this custom GPT won’t be needed much longer. One day, ChatGPT should just pick the best model for you.

Try the AI Model Advisor for GTM. See link in the comments. Below is a 90-second demo video of the GPT in action. Let me know your thoughts.

AI Model Advisor for GTM (Custom GPT): https://chatgpt.com/g/g-684c1e4539448191be08b3a7e26a5361-ai-model-advisor-for-gtm

The AI industry has a language problem. And it’s unintentionally holding people back.

I work with GTM teams to show them what’s possible with AI through real, applied use cases. Some of the brightest, most capable people I know hold back because the language sounds intimidating. Like it’s designed for certain types of people, but not them.

Generative AI is one of the least technical technologies we’ve ever seen. You talk to it like you’re talking to a colleague.

Yet we’ve wrapped it in jargon:

➡︎ Multimodal AI

➡︎ Context window

➡︎ Retrieval-Augmented Generation

This creates real issues. The sales rep who could write better follow-ups doesn’t try. The marketer who could generate new campaign ideas is stuck with what they’ve done before. The customer success manager who could draft empathetic responses in seconds keeps doing it manually.

Words matter when we want AI to benefit all of humanity, not just those comfortable with technical jargon.

Below is a fun 30-sec quiz that I “vibe coded” into existence using AI to prove this point. 🤣 See link in the comments.

If you can explain something simply, you understand it well. If you can have a conversation, you can use AI.

To AI companies: you can do better. Start with the language.

How did you do on the quiz?

Take the 30-sec quiz and see how well you cut through the BS: https://claude.ai/public/artifacts/76659222-505e-4774-81af-644150b292e2