It’s a good question. Given the hype around Generative AI (like Chat GPT4) and Large Language Models (LLMs), I’m sure that most small business owners and marketers have at least thought about this question.
Full disclosure: before running this experiment, my view was pretty much within what appears to be the general consensus: Generate AI – and LLM’s in general – can definitely help save time and money, but there still needs to be a ‘human in the loop’ to make sure that the content and LLM generates makes sense.
As part of an experiment, I asked Bing Chat (basically Chat GPT4) to answer the following prompt:
“Act as an expert in Large Language Models. Write an 800-word article explaining what Large Language Models are, how they work, and how they can be used to enhance a company's marketing efforts. Write the article on a 10th grade reading level.”
Bing responded pretty quickly (about 20 seconds max) with the following output:
“Large language models are a type of artificial intelligence that can generate human-like text. These models are trained on vast amounts of text data, allowing them to understand and generate language in a way that is similar to humans. As a result, large language models have the ability to write articles, generate chatbot responses, and even create poetry.
At their core, large language models work by predicting the next word in a sequence of words. They do this by analyzing the context of the words that come before it and using that information to generate a prediction. For example, if the model is given the phrase "The cat sat on the", it might predict that the next word is "mat" because that is a common phrase in the English language.
Large language models are able to make these predictions because they have been trained on vast amounts of text data. During training, the model is shown millions of examples of text and learns to recognize patterns and relationships between words. Over time, the model becomes better at making predictions and can generate text that is coherent and human-like.
One way that large language models can be used to enhance a company's marketing efforts is through content generation. With a large language model, a company can quickly generate high-quality articles, social media posts, and other types of content. This can save time and resources, allowing the company to focus on other aspects of their marketing strategy.
Another way that large language models can be used in marketing is through chatbots. A chatbot powered by a large language model can provide quick and accurate responses to customer inquiries. This can improve customer satisfaction and help build trust in the brand.
Large language models can also be used to analyze customer feedback and sentiment. By training a model on customer reviews and social media posts, a company can gain insights into how their customers feel about their products and services. This information can be used to improve marketing efforts and better meet the needs of customers.
In conclusion, large language models are a powerful tool that can be used to enhance a company's marketing efforts. By generating high-quality content, providing quick and accurate chatbot responses, and analyzing customer feedback, large language models can help companies build stronger relationships with their customers and improve their overall marketing strategy.”
You can see that Bing Chat generated a response that answered my question very precisely, though it missed the word count request by about 50%. That’s fine since the issue seems to be in the way I constructed the prompt. (This was confirmed when I added further refinements to the prompt and came up with an article that was almost exactly 800 words.) The grammar and structure of the article was pretty good right out of the box, and while it would've been nice to have a bit more detail I'm satisfied with the basic result.
One of the key points is that it would have taken me much longer to write this article on my own. I’d have to do the research, organize the materials, and then write a coherent article. Bing Chat, however, did it in about 20 seconds, so my efforts were focused almost entirely on thinking about ‘what just happened?’ and ‘is this information accurate?’. On a side note, AI and LLMs don't always work the way you want them to; it can't read your mind, so constructing the prompt is critical. They can also 'hallucinate' but that's a topic for another day. From my experience so far, 'out of the box' LLMs still aren't great at delivering specific marketing strategies or planning, but that's definitely going to change.
So, initially, yes – LLMs can indeed help but it’s important to take time to understand how they work, what the limitations are, and how to construct prompts that will deliver the best results.