Confessions of a Chronic AI Chatbot Hopper
Key Highlights
- Prompt engineering is key; asking AI for help crafting prompts improves output quality across platforms.
- Running similar prompts through various chatbots allows users to evaluate consistency and uncover diverse insights.
- Using AI to analyze, review and optimize content enhances visibility and strategic decision-making in digital marketing.
If my AI usage were compared to a TV show, I’d fit snugly somewhere between "Sex and the City" and "The Golden Girls." Specifically, I’d be the AI-using equivalent of a Samantha or a Blanche.
I’m not exactly sure how many interactions I’ve had with AI chatbots — I use them every day for various tasks — but I’m noncommittal to them. Whereas Alexis Gajewski writes about being a one-chatbot kind of lady, I tend to date around (or whatever the AI equivalent is). I’m pragmatic; I’m often pitting one chatbot against another to see which one produces the outputs I prefer.
I also tend to be a woman of mystery when it comes to my chatbots. None of them really know my life story or what I like or dislike. I once named one of my ChatGPT accounts, but I often forget what I called it, so we’re on a no-name basis now.
I’m not a reckless user, nor do I disregard AI, but I do like having choices and want to find the best available answer.
Move over Cyrano de Bergerac, Claude de Gemini is taking over
Alexis has been my AI prompt teacher for about a year now. She taught me something last year that has been invaluable as I hop, skip and jump from one AI chatbot to another: have AI help you use AI better. Essentially, it means having one AI chatbot help you put together a prompt that you can use on another.
My specialty at Endeavor is visibility — basically, how to show up in both search engines and answer engines. Affectionately known as “The Optimizer,” I’m surrounded by mountains of search data and traffic metrics and am sometimes at odds about how best to analyze everything in front of me.
Rather than turning my brain into mush trying to figure out how to craft the perfect prompt for Claude, I turn to ChatGPT and ask it the best way to talk to Claude. (I don’t refer to them by name in the prompts. I’m not a monster!)
For example, I’ve been known to ask ChatGPT things like:
- What’s the best way to write a brand analysis prompt?
- What prompts will help conduct a comprehensive competitive analysis in generative engines?
- You're analyzing the backlink profile of a website. What are the best prompts to use to create an exhaustive review of backlinks?
- You’re an AI adviser and an SEO analyst tasked with developing a spreadsheet of topics for a content calendar. You’re interested in determining the highest priority topics based on traffic data, including URL, views, visits and bounce rate. Write a prompt I can use to extract this information into a spreadsheet and put it in order of highest priority. In the prompt, please also include the reasoning for this ranking
Based on what prompts ChatGPT gives me, I tweak them to my needs and then run them through whatever AI chatbot will be most useful for the task.
50 first prompts
I also use multiple chatbots to run similar analyses to see whether they produce the same outputs. I’m not sure what it says about me that I thoroughly enjoy asking Claude, Gemini and ChatGPT to review my copy for syntax, clarity, grammar and punctuation — and then watching to see whether they’ve all clocked the same errors. Because of how large language models are trained, most of them do.
I also run similar prompts through each LLM to see what they tell me, and how they tell it to me. Will Gemini woo me with its Google intel? Will Claude continue to show me how it thinks, effectively pulling me closer to see what’s behind its vast data sets? Only time will tell, but I assure you, I often sit with bated breath as I watch what each chatbot tells me about the questions I ask.
Speaking of questions, I often ask my stable of eligible chatbots what they think of me — or at least the brands I’m working on. It’s like the modern-day equivalent of a friend asking the person you like whether they like you back and if they’d go out with you. Don’t be fooled: It’s not all flowery language and love. I’ve definitely felt the gut punch of a chatbot telling me it didn’t see the inner beauty of the seven-part explainer series I put together. Similarly, hearing one chatbot tell me that my "About Us" page could use considerable work left me feeling rejected.
Sliding keyboards
All of this is to say, if my AI habits make me the Samantha or Blanche of AI users, I’ll take it. I’m experienced enough to know what I want, confident enough to test my options, and practical enough not to mistake some convenient attention for quality. My relationship with AI is intentional, not emotional. I ask better questions, compare responses, and I have the final say.
Like what you're reading? Get MarketingEDGE delivered to your inbox every week!
About the Author

Erin Hallstrom
Contributor
Erin Hallstrom is the Director of Content Operations and Visibility for EndeavorB2B, where she works with more than 150 trade journalists across 90+ brands to implement search engine optimization (SEO) and generative engine optimization (GEO) best practices. She’s been a featured speaker at the News and Editorial SEO Summit (NESS) and headlines ASBPE’s SEO for B2B Media Playbook education series.
In addition to optimization strategy, Erin is responsible for Endeavor's metrics reporting, where she uses her expertise in website analytics to help teams understand their data to make informed content decisions. Erin holds multiple technical certifications in Google Analytics and also trains audience and marketing groups how best to utilize SEO and GEO tactics for enhanced content marketing performance.
Resources
Quiz
Elevate your strategy with weekly insights from marketing leaders who are redefining engagement and growth. From campaign best practices to creative innovation and data-driven trends, MarketingEDGE delivers the ideas and inspiration you need to outperform your competition.

