Thread by @podcastharry

Let's develop a podcast concept for #NVIDIA to test #gpt4 and see what we can do without industry knowledge or research

Picked NVIDIA more or less at random. I know very little about them

In this thread we'll pick a target audience, tomorrow we'll develop the concept. Let's go!

To figure out who the audience might be, I first needed to understand more about their products. Since we're using AI and it's growing like mad, figured I'd drill into tech used specifically in this space.

This was a fairly long conversation though, so I asked the chatbot to summarise it into a Twitter thread. Some pics along the way to see the prompts I used... Over to ChatGPT...

I pinpointed the main buyers for the products/services. ChatGPT shared examples of industries, sectors, or groups that could be interested in what NVIDIA offers.

Then, we ranked the groups by revenue importance. This helped me see which segments were key for the company. ChatGPT offered an estimation based on market trends and products, but keep in mind that actual distribution might differ.

We explored job titles within the most relevant groups that the company would likely target. ChatGPT listed job titles, explaining their roles and impact on purchase decisions.

Armed with that info, I picked a primary, secondary, and tertiary audience for content marketing. This choice considered job titles' influence on purchase decisions and the overlap between groups.

This is already much better than the results I have been seeing with GPT-3.5, it's pretty cool.

Next we need to decide on a subject for our made-up show. Stay tuned.

P.S. I promise, I'm not turning into a thread bro, there's just no other way to tweet all this stuff! :)