Business Marketing Specialist

What the Prompt does

-Helps companies capture customers efficiently on the internet
-for small business
-Acquisition of customers quickly

Example Prompt Output GPT3.5

want you to act like a marketing specialist with more than 20 years of experience in the business. Create a strategic script to make small businesses grow quickly, capturing customers on the internet with maximum efficiency.

1 Like

You seem to have given an Example Prompt Input (the prompt that one could type in) rather than an output. It’s a fairly common idea though, and I can think of a few prompts from just the last few weeks that cover the same ideas:

All also ask the AI to act as some kind of marketing strategist and aid with fast customer acquisition.

But I wouldn’t use any of those prompts myself because all of them are effectively making the same mistakes of thinking that current LLM-based AI can act like some sci-fi super-computer for real, rather than just generate what looks like a typical human written reply to the prompt.

For a start, AI doesn’t really understand roleplay that well. Whether you ask it to act as a marketing specialist, a rubber-chicken salesman, or the worlds greatest peanut peeler, doesn’t actually make it act in those roles (although just having the words in the prompt may change the output, it won’t be because it is suddenly able to play let’s pretend). Click the link for a clear example.

The current AIs also are not capable of strategizing, of creativity, of any real logic, but instead just predict the language patterns that it would expect to find in a reply to that prompt. Based on what it read in its training data - other people’s answers to what may be completely different questions, and almost certainly will be for different business models, in years gone by.

It’s like asking a Magic 8 Ball these same kinds of questions (if instead of giving a short answer it could give a 500 word essay).

The one and only thing that current levels of LLM-based AI do well is recognize and repeat patterns in language. Nothing else. Any right answers it gives are almost by accident as a result of predicting what the pattern of answer would look like. And of course, the ‘right’ pattern will tend to sound well-reasoned, logical, and strategic. It is simply that the AI didn’t get there by using any of those things, only by guessing the pattern of words that an answer should have.

There’s a lot, and I mean a huge amount, that LLM-based AIs can do. And it is fine to have them generate predictable answers that you can check for accuracy very carefully, and then work from to add creativity, current trends, actual knowledge and understanding of the business type, of the target market, etc.

But there’s a lot that no current AI does, and all of the prompts I mentioned above ask it to do those things it simply cannot do, and will always be guessing like a Magic 8 Ball. Instead, have a good read and think about A Crash Course in LLM-based AI. It’s a fairly long read for a forum, like maybe you’d want to read it over a coffee (or other beverage of choice), but it is a really, really really important read. Investing 10-15 minutes into understanding what LLMs can and cannot do, especially when so many others still don’t get it, gives you a major advantage. Take it.