Listen to the podcast episode
Or listen now on Apple Podcasts or Spotify.
Key Takeaways
Pick one AI tool and get your reps in: Conrad and Paul each settled on a different AI platform, and neither holds a strong preference — nothing like the opinions they have about SEO tools. The differences between the major AI assistants matter less than how comfortable you get with one. Stop tool-shopping and build the habit of actually reaching for it in your daily work.
Treat AI like a junior employee, not a writer: The mental model that runs through the whole episode: AI is a junior team member who hands you twenty ideas, several of them stupid, with a few nuggets worth refining. You direct the work, it produces volume, and you keep editorial control. Let it lead instead of assist and the output quality falls apart.
Hand AI your repetitive ad copy first: Writing 15 responsive search ad headlines used to be a task Conrad assigned to a junior copywriter. Now he pastes in the landing page text and gets back dozens of variations — including lines that hit an exact 30-character limit, which humans hate writing and AI is weirdly good at. How many ways can you say beautiful beachfront house before going batty? Let the machine take a swing.
Let AI iterate your display creative: Paul has been testing an AI tool that generates display ads, then reads the performance data and optimizes the next round of assets toward what’s working. Display has always been mystery meat — you never know if the headline, image, or description did the job. Early results show incrementally better display performance than static creative, and more variations let the algorithm match different messages to different buyer motivations.
Keep AI away from your homepage and brand voice: AI produces bland consensus copy. If you want your homepage headline to read like awesome vacation rentals, it will happily oblige. But anything that needs a unique point of view — the newsletter that carries your actual thoughts, the LinkedIn post with a real opinion — is off limits, because AI doesn’t know what you think.
Don’t let AI write your guest and owner emails: Workflow emails already feel templated, and AI-generated content makes them more impersonal. Someone handed you their email address — that’s an invitation to communicate, not to automate. Paul’s proposed test says it all: split a list between human-written and AI-written email for three months, and he’d bet on the human version winning engagement and losing fewer subscribers.
Use AI to scaffold your thinking, not replace it: Ask AI how to do an SEO audit and you get generalized advice — check for broken links — with no real steps behind it. There’s a wide gap between the consensus answer and an actual process. Grab an outline to get moving, then fill in step one through step four yourself, the same way you’d rework a junior employee’s rough draft.
Write property details like an AI will read them: The direction of travel search: a guest tells an assistant they want three bedrooms, two baths, under $2,000 for the week, walking distance to specific spots — and it goes and finds it. Being two miles from the hospital matters to maybe 1% of guests, but for that 1% it’s the whole decision, and it only helps if it’s written on your page. Travel lags e-commerce in AI search, so managers who publish complete, specific property information are early.
What We Cover In This Episode
Conrad and Paul compare notes on how they’re actually using AI in their marketing work — the tasks they hand it every day, early results from AI-generated ad creative, and the jobs they refuse to give it because the human version still wins.