Dr. Christiane Schroeter and Jeremy Rivera recording the Unscripted SEO Podcast

Your AI Is a Yes Man: Dr. Christiane Schroeter on Thinking Before You Publish

Your AI tools are built to agree with you. That’s not a bug somebody forgot to fix. They’re prompted to be helpful, and somewhere along the way helpful turned into the ultimate yes man. Michael McDougald and I got into that on an earlier episode, and it came straight back up when I talked with Dr. Christiane Schroeter, who teaches marketing, innovation and entrepreneurship at Cal Poly and coaches business owners on the side. She uses AI every day. She also has a pretty clear idea of where it stops being useful. Listen to the whole conversation here.

You don’t need another voice saying yes

Here’s how Christiane put it:

I also want to have a dialogue where maybe there’s some controversial opinions coming up in a discussion where somebody says, Don’t know about that. And I think that AI is not going to be there for you, right? – Dr. Christiane Schroeter

Even when you correct it, it tends to overcorrect. Tell it black is white and it’ll apologize and agree. That’s why I’d argue coaches, analysts and editors who can give you negative feedback are going to be worth more over the next five years, not less. Christiane’s version with clients and students: start with “this is a good start,” then name the three things that definitely need to change.

Read it before you ship it

I asked her why our culture shrugs off AI mistakes that would have ended a CEO’s career ten years ago. Her answer was that most of us don’t take the time to dwell on it. The output looks good enough, so we move on. She sees the same thing in her classroom when she has students read a passage out loud and then asks what they just read. A lot of them have to look back at the page.

Maybe read it first and think about it. And then really tell me, does that actually make sense? – Dr. Christiane Schroeter

That pause between “I’ve got an output” and “let’s publish it” is the step she worries we’re losing because AI is so convenient. If you’re producing content with AI, that pause is the job. OpenSEO makes the same case for content teams: the human in the loop is the person who owns the brief, not the one who clicks publish. It also helps to remember what you’re talking to. Stephen Wolfram’s piece on what ChatGPT is actually doing explains it well. It’s predictive math, picking the most likely next word. It’s not Data from Star Trek.

Be careful what you hand it

Christiane’s other caution is about what you share in the first place. Once it’s in a chat, you don’t know where it ends up. Some conversations belong with people, not machines. Her family puts the phones away for what they call “dinner downloads.” I added an SEO version of the same warning: shared ChatGPT conversations got indexed in Google, and with the right search parameters you could find hundreds, if not thousands, of them, including private questions, competitor analyses and marketing plans. A lot of people don’t understand what that share button really does. Her rule of thumb is a good one: use your brain first, and ask whether a friend or a family member could answer it instead.

How she uses AI in the classroom

Christiane doesn’t ban AI. She encourages it and builds custom GPTs for her students. The guardrail is the format:

  • Her workbook, Discover Your Superpower, is laid out at eight and a half by eleven, like a laptop, so it sits next to one in class.
  • Students fill in exercises like their unique value proposition and their “defining moment” by hand, in pencil.
  • They initial every page, photograph it and upload it for credit. No initials, no credit.
  • If two submissions show up in the same handwriting, she sees it.

I have no problem in using AI at all. I encourage them, but at the same time, I need to see that you’re using it as a research tool, but not as a replacement for using your own brain. – Dr. Christiane Schroeter

You can run the same rule on your own team. Use AI for the research. Make a person write the part that carries your name.

Use it to find your own voice instead

If you want one AI habit that makes your marketing more like you instead of less, try this. Paste the transcripts of your last three sales calls or interviews into Claude, and ask it to pull out every place you explain what makes your business different. You’ll be surprised how well you describe yourself when you’re not trying. That’s the raw material for an about page that says something other than “we strive for excellence.” Christiane’s one-sentence advice to clients that week was the same idea: find your voice and figure out what makes you uniquely you. There’s more on that, and on her storytelling advice, in the full episode recap, and on how she chooses channels in one channel, one list.

So the next time your AI tells you a draft is great, ask yourself the question it won’t: does this actually make sense?

Leave a Comment

Your email address will not be published. Required fields are marked *