AI For Content: The Truth About Large Language Models (LLMs) in Marketing
Here’s the problem: AI for content is commoditized. It’s hyped, it’s packaged, it’s sold as a miracle that will solve the chronic understaffing that has plagued this practice since its inception. It will not.
How do I know? I’ve been testing, re-testing, honing prompts, pitting platforms against each other, and sitting through interminable platform demonstrations to uncover the truth of what Large Language Models (LLMs) can do, and what they can’t.
Why LLMs Alone Won’t Make You a Better Marketer
That truth is constantly shifting and evolving. Which is why a multi-year testing strategy is required to even attempt to keep up to date on what technology is leading the playing field.
For the past two years, I have been testing different LLM platforms for different use cases, and here is the truth as I understand it in 2024: LLMs can make you more efficient, but they can’t make you a better marketer.
Practical AI Insights for Small Business Owners
I recently presented my findings to a group of female small business owners to help them integrate AI into their marketing efforts.
There were no enterprise-scale case studies or big ideas about eliminating manual workflows. Instead, I explained:
- How to select the best platform for their purpose.
- How to manipulate prompting to get the best results fastest.
- When to consider a paid option (hint: not for a long time).
Actionable AI Insights for Marketers
My research has led me to the following conclusions, which I implement in my role as Director of Content Marketing, and share with other women interested in the intersection of LLMs, marketing, and creativity:
The Right AI Developer Matters
- Anthropic/Claude.ai has a much higher standard for privacy out of the gate than any other LLMs.
- I can see a future where Google penalizes content developed with LLMs, with the glaring exception of Gemini-generated content.
- Wu Dao might be further along than any of the rest, but considering the TikTok drama, an LLM built in Beijing could easily get banned in the US.
Know thy LLM developer, know thyself.
No Single AI Platform is Best
There’s no top gun. Different platforms have different strengths and weaknesses, and it’s constantly shifting.
Line in the sand right now:
- Claude.ai is the best at human-sounding copywriting, but it delivers complete falsehoods when asked for statistics.
- Gemini sources information for you but will use the word “epitome” 15 times in a paragraph.
- Copy.ai allows you to dictate a brand voice, but strangely, all those voices sound the same.
Anyone (friend, foe, SaaS rep) who tells you one is the best is selling something.
Unlock the Power of AI Tools with Prompt Engineering
You are the creative force. Prompt engineering is a new language we all have to learn to harness this technology.
If you don’t know how to ask for what you need, you’re not going to get it. The biggest hurdle to adoption I see is people entering a thin prompt, not liking the result, and giving up.
That’s like asking one question on a first date and then walking away. Prompting is a conversation, and prompt refinement is the art of asking the right questions until you get what you need.
Pro Tips for Mastering LLM Prompting:
- Prompt engineering is a skill, and mastering it is essential.
- Thin, generic prompts yield poor results.
- Refining your prompts through iterative conversation is how you unlock real value.
- If you’re frustrated with early results, don’t give up—refine your approach.
Separating Valuable Tools from Hype
This is our generation’s Gold Rush. Everyone wants a piece, the barrier to entry is low, and a lot of people in California who don’t know what they are doing are going to fail.
Some platforms legitimately gather multiple LLMs into one interface and build upon them to provide a valuable, scalable content production tool. The problem is they present the same as a basic GPT wrapper, and it takes valuable time to distinguish who is the real deal and who is pedaling Fool’s Gold (cough, Ryter, cough).
Too many content departments are being pressured to “get in the game” and utilize a platform when it’s smarter to master the free tools and wait for capitalism to bring the best solution to the top of the pile.
Debunking Misconceptions About AI in Marketing
I don’t know if anyone is materially better off from my research and advocacy. I do know that I’ve helped businesswomen I admire overcome some initial barriers to engaging with this technology.
I do know that if people are burned out, wildly under-resourced, or simply weren’t that dedicated to content marketing in the first place, ChatGPT will not save them.
Correcting that misconception in the marketing industry is important to me.
Embrace AI Tools Without Losing Sight of the Goal
We’re here to tell stories. If Gemini can help prompt a new idea, use it! If you need to draw on a napkin to demonstrate your point, scribble away. If you can only express an idea through movement, film a video.
These are all tools to communicate, and all are equally useless if you don’t have something to say.
If you want to learn more from our marketing experts, reach out below!