What Marketers Get Wrong About ChatGPT: Lessons from Britney Muller

Britney Muller is one of the few people I trust when it comes to AI. Not because she hypes it up, but because she understands what’s actually happening behind the scenes when you ask ChatGPT a question. She’s been deep in machine learning since 2012 (back when she found a Harvard data science course and consumed it “like a favorite sitcom”), and she has a way of explaining complex concepts that makes them accessible without dumbing them down.

The most effective way to use ChatGPT for marketing is to start with ONE specific, repeatable task, such as summarizing meeting notes or drafting email outlines, and master it before expanding. Treat ChatGPT like an eager but sometimes wrong intern: verify important outputs, give specific instructions, and build trust gradually.

I wanted to have this conversation because when we bring on new clients at Kick Point, we often hear the same thing: “We know we should be using AI, but we don’t know where to start.” And honestly most of the advice out there isn’t helping. It’s either too basic or so advanced it feels out of reach.

Britney’s approach is different. She gives her students permission to slow down, get specific, and treat AI as a tool that enhances their skills, not one that replaces them. Liz on our team took Britney’s course in late 2025, and the things she learned were incredible.

Let me share the frameworks from our conversation that I think will change how you approach ChatGPT.

This resonated with me because I see the same pattern with analytics implementations.

The biggest mistake marketers make with ChatGPT is trying to do everything at once. They sign up for every AI tool, try to automate their entire workflow, get overwhelmed, and eventually conclude that AI “doesn’t work for them.”

Instead, choose something you do repeatedly. Summarizing calls. Writing briefs. Editing drafts. Research for a specific type of project.

Get genuinely good at using ChatGPT (or whatever AI you like!) for that ONE thing. Build confidence and understanding before adding more use cases. This takes some work, but the payoff is that you develop real intuition for what AI does well and where it falls short.

Britney tracks every AI interaction to learn what produces good results and what doesn’t. She treats it like an experiment, not a magic solution.

This is when ChatGPT for marketers becomes genuinely powerful because you’ve built enough experience to know when to trust it and when to verify.

I hear the fear a lot: “But I’ll fall behind if I don’t move fast.” In my experience, the teams who try to do everything all at once usually end up doing nothing well. Starting small isn’t falling behind! Think of it as being strategic about where you invest your learning time.

ChatGPT is like a particularly spicy intern. We’ve all met one (or were one), right? Eager, confident, and sometimes completely wrong. Using this mental model is incredibly helpful for setting realistic expectations about how to use AI in marketing.

What makes an intern “spicy”:

They’re enthusiastic and confident (sometimes overconfident). They’ll give you an answer to anything you ask, even if they don’t know the answer. They need guidance, feedback, and oversight. And they get better with clear direction and specific tasks.

Why ChatGPT fits this description:

Here’s how Britney put it: “AI has no ground truth. It doesn’t live in the world we live in. It doesn’t know what gravity is. It’s trained on all of the world’s text. AI has no friends. It has no strong opinions.”

This results in AI sounding confident while being completely wrong. The confidence in AI responses doesn’t correlate with accuracy, and that’s where marketers can get into trouble.

What this means for your workflow:

You might consider building verification steps into your process as a standard practice, not an afterthought. For anything involving facts, figures, or recommendations, I’ve found it helpful to verify before publishing or sharing.

Give ChatGPT specific, well-defined tasks rather than vague requests. “Help me with marketing” produces generic slop. “Review this email for clarity and suggest three alternative subject lines under 50 characters” produces something useful.

Treat every output as a draft, not a final product. This can be a tough habit to break — especially when ChatGPT sounds so confident! But I’ve found that building in verification steps makes AI more useful, not less, because you start trusting it for the right things.

I think this reframe addresses a fear a lot of marketers have.

The most valuable way to use ChatGPT isn’t to outsource your thinking. Rather, it’s to enhance your existing skills. Britney uses the metaphor of “sharpening the axe.” You’re using a tool to make yourself more effective, not handing your axe to someone else.

Here are three approaches I’ve found helpful:

Use AI for editing, not creating from scratch

Start with your own draft (even a bad one!) then let AI help refine it. You maintain your voice and expertise; ChatGPT helps with structure and polish.

This approach is often faster AND produces better results than starting with AI output and trying to make it sound human. Your expertise is the foundation.

Use ChatGPT to identify gaps in your knowledge

Ask ChatGPT to critique your work or identify what you might be missing. Use it as a research assistant to explore adjacent topics. Let it surface questions you hadn’t thought to ask.

This is especially valuable for ChatGPT for market research, such as using it to map out a topic landscape before you dive into primary research.

Build custom GPTs for repeatable workflows

Once you’ve mastered a specific use case, consider systematizing it. Create custom instructions that encode your preferences and standards. This compounds your efficiency over time and makes your ChatGPT prompts for marketers more consistent.

The bigger picture here is encouraging: people who learn to use these tools effectively will be more valuable, not less. The skill isn’t “using ChatGPT”. Instead, it’s knowing when and how to apply it to your specific work.

Britney made a prediction that I sincerely hope comes true. She said that after the wave of AI-generated content (what we’re all calling “slop”), there’s going to be a renewed hunger for authentically human work.

“We’re sick of this. We crave real human content. We crave real photos. We crave real personality filled writing.”

The upside of the flood of mediocre AI content creates an opportunity for differentiation. If everyone’s producing the same ChatGPT-generated listicles, what happens is that genuine expertise and authentic voices become more valuable.

What this means for you is that the skills that make you valuable such as creativity, judgment, and your authentic voice aren’t going away. In fact, they’re becoming more important as AI makes generic content cheaper to produce. Marketers who have these skills will stay around.

Remember: the “spicy intern” can help you work faster, but it can’t replace what you bring to the table: lived experience, relationships, and the ability to make judgment calls about what matters.

As people adopt ChatGPT, AI Mode, and other LLM generated responses, it creates new complexity for how we measure and attribute results. If you’re using ChatGPT to help create content, how do you measure the impact of the human contribution versus the AI contribution? AI-assisted workflows make it harder to isolate what’s actually driving results.

What you don’t want to do is focus on the AI “efficiency gains” and ignore those quality outcomes. Sure, you’re producing content faster, but is that content performing better? Sometimes yes, sometimes no and without measurement, you won’t know which.

Is the “spicy intern” actually saving you time? A piece of content that takes 30 minutes to generate but requires an hour of fact-checking and revision isn’t necessarily more efficient than one that took 90 minutes to write well the first time.

These are questions we’re all asking ourselves and I don’t think anyone has all the answers yet. But I think starting with clear metrics for what “success” looks like, well before you implement AI tools, will make a big difference in whether you can honestly evaluate their impact.

1. Watch the full webinar replay

You can find our conversation here on YouTube. We covered a lot more ground than I could include here, including some great audience questions! The full video is about an hour long.

2. Check out Britney’s resources

Britney has incredible resources for going deeper on AI:

• Her course: Actionable AI for Marketers on Maven (Liz took this and loved it!)
• Her community: Orange Labs
• Her website: britneymuller.com

3. Try one thing this week

You might consider starting with just one approach:

• If you’re feeling overwhelmed: Pick ONE repeatable task and master ChatGPT (or your favorite AI tool) at doing that task before expanding
• If you’re skeptical: Try treating ChatGPT like a “spicy intern” and see how verification changes your trust level
• If you’re worried about job security: Focus on using ChatGPT to sharpen your existing skills, not replace them

Just start somewhere! You won’t transform your AI approach overnight, but one step at a time gets you there.

I learned a ton from this conversation with Britney, and I hope you found it valuable as well. Let me know in the comments of the YouTube replay what resonated most with you.