Britney Muller is on a mission to empower marketers through AI education and to shed light on the misinformation and hype surrounding AI. In this Whiteboard Friday Revisited, we look back at her 2020 episode on "Accessible Machine Learning for SEOs." Join us as we revisit the past to understand the present and speculate on the future.

Five years after this video was created, the idea that machine learning might be “far outside the scope of your SEO work” seems outrageous. But you can understand why it was positioned this way. Britney’s early work in LLMs, machine learning, and SEO was initially met with scepticism. 

"I remember people telling me I was insane and actually, like, very seriously criticizing me for making these sorts of leaps when I have been at the hackathons. I've been at these AI conferences. I know the current capabilities exist," but at the time, the concept was "such a foreign concept" to many people.

As we enter the second half of 2025, three years after the ChatGPT general release, there is no shortage of AI marketing workflows, including on our very own Moz Blog. 

But even Britney admits she couldn’t have anticipated that LLMs would be the model that would "explode in the way that it has" with the wide release and adoption of tools like ChatGPT. Instead, she thought that predictive models or data models focused on "analysis and surfacing statistical insights" would take off first, as those applications had "incredible capabilities through other types of machine learning" that are now being overlooked. Perhaps this just proves that predicting anything in this crazy world is nearly impossible. The real lesson? Stay informed so you can react in the most sensible way possible. One great way to do that is through Britney's Maven Course, Actionable AI for Marketers. 

Britney Muller advises that adding a "human in the loop" is crucial when using AI tools for anything, but especially content. This human touch involves refining and adding a layer of expertise and personality to the AI's output. She emphasizes that a person's unique expertise, real-world experiences, and anecdotes are things AI can never replicate. For a good example, Britney references Ed Zitron's blog as content that is both "highly informative and entertaining" with a "unique voice" and "humor throughout that is so uniquely human." While this piece was published after recording this interview, it is a good example in tone and topic. Does all this sound familiar? That's because you know this already, from our good, old friend EEAT. 

It all starts with clarity of thought and a mindset shift. Britney emphasizes that to use AI effectively, you have to be willing to embrace failure because it's the "yellow brick road of using AI." The process is inherently experimental and should be treated as such. She advises thinking like an engineer, which means breaking down problems into small, manageable chunks and being specific in your prompts. Avoid the "magic engine" mentality where you provide a vague prompt and expect a perfect result.

You can begin by documenting the tasks that take up a lot of your time or are painful to do. This helps you identify which workflows could be most effectively automated with AI. Once you've honed in on a task, select the right model for it. While generative content is what people often think of, Britney suggests it's actually one of AI's "worst capabilities." Instead, explore using AI for more efficient tasks, such as:

• Data analysis: Analyze large datasets to surface unique insights.
• Automating mundane tasks: Use AI to identify qualified leads on social media.
• Building small tools: Use models like Claude to help create simple Chrome extensions.
• Running local models: Explore models like Llama, which can run directly on your computer to query against your own documents and create a mini RAG model.

Because this is, after all, a marketing blog, I wouldn’t leave you without a framework, so here it is:

• Reframe your approach: Don't get discouraged by AI's experimental nature and the failures that come with it. Embrace the fact that failure is the "yellow brick road of using AI".
• Start small and be specific: To avoid being overwhelmed, begin with small, simple tasks. The goal is to find one clear workflow that solves a problem for you.
• Learn to think like an engineer: To use AI effectively, you need to break down problems into individual, actionable chunks, as you would when instructing a person. I also like to think of this as being incredibly meta, or thinking of how you’d explain this to a junior employee (sorry, job-seeking graduates 😞).
• Go beyond content generation: While generative AI is good for some tasks, it's not well-suited for creating high-quality content. Explore other applications, like using AI to analyze data sets, check for grammar in emails, or set up automations to identify qualified leads on social media.
• Use the right tools: Not all AI models are the same. For instance, a text-to-image generator like Midjourney is different from the computer vision used by the post office to read addresses. You can even run your own local AI models for specific tasks on your computer.
• Document your workflows: As you experiment, document your processes. Tracking time-consuming tasks can help you find opportunities for automation.

According to Britney, you're not to be blamed for getting stuck in a hype-induced fear, hopelessness, urgency cycle. 

"There's so much misinformation happening right now in terms of AI and the AI hype that circulates at an incredible speed and I like to shed light and give context around why these things are occurring and also give marketers peace of mind that they're not going to be replaced by AI tomorrow."

The real tragedy is the misuse of models clouding the real benefits of AI by orchestrated LLM hype and the muddying of the world's information with "AI slop." As marketers, we're perpetual optimizers, and the temptation to let a "mountain of AI slop" flow to "keep up" is a powerful one, but the return on this is likely short-lived and will result in diminishing returns.

This brings us back to the recurring SEO is dead narrative. As many of us know, touted new SEO/GEO/etc methods are often just a rehash of existing, timeless strategies. The key is to be informed and proactive. A constructive path forward is possible without succumbing to the fear of being left behind.