Show Notes: Inside the 90™ Episode #40
Part one of this AI conversation covered vibe coding. This second half moves into everyday LLM use — and a newer term worth knowing, vibe marketing.
Same caveat as before: this is two entrepreneurs sharing what's actually working, not a claim to expertise in AI strategy or governance.
Building a Coaching Co-Pilot With Custom GPTs
A custom GPT loaded with real team context can turn a vague feeling that a conversation needs to happen into an actual, well-prepared one.
A quick platform note first: custom GPTs, not the newer "projects" feature, are the more secure option for anything sensitive, since they don't feed back into the underlying model. That distinction matters once you start loading one with real team data.
One especially useful build is a dedicated coaching GPT, fed with each team member's Kolbe scores, Working Genius results, and the company's core values and accountability chart. LLMs don't automatically parse something like a Kolbe score correctly, so each instinct has to be spelled out explicitly rather than pasted in as a shorthand number.
For a leader who's naturally more objective than empathetic, or simply moves faster than the person they're coaching, this becomes a genuine rehearsal space — asking how to deliver feedback to someone with very different instincts, and getting back thoughtful, specific questions rather than a canned script.
Enhancing the Conversation, Not Replacing It
The test for whether an AI use case is healthy: is it enhancing a conversation, or quietly replacing it?
Using a GPT to prepare for a tough one-on-one is enhancing. Using one to write a finished rock so a team can skip the discussion is replacing — and it costs the team the actual alignment that writing a rock together was supposed to produce in the first place.
This maps directly onto EOS's leadership and management principles. Leadership means giving clear direction and creating the opening for people to engage; nowhere does it say the leader has to arrive with the answer already solved.
Used well, AI can actually support that letting-go instead of undermining it — helping someone lead with the next logical step instead of manufacturing false clarity about having it all figured out.
The real threat to good communication usually isn't the tool — it's impatience. The goal is a better conversation, not a faster one.
Protecting the Human Moments
The healthiest rule found so far: AI stays completely out of the room during L10s and quarterly planning.
Those stay pen-and-paper, human-only sessions, on purpose. Alignment and clarity are the actual product of that time together, and no tool should shortcut the work of getting there as a team.
Outside of those protected sessions, the cultural impact of encouraging AI experimentation has been almost entirely positive and mostly unplanned. Making an AI-related rock part of the regular quarterly rock rotation has produced real enthusiasm rather than anxiety about job security — everyone's finding something that adds value to their own role.
Vibe Marketing: From SEO to AEO
Search hasn't shrunk — it's split. Total search volume is up, but a meaningful and growing share of it now happens inside an LLM instead of a search engine.
That shift has a name: AEO, or answer engine optimization, alongside SEO. The good news is that a genuinely strong SEO strategy was already most of the way to a strong AEO strategy.
The core move for both has always been the same: write content that answers the real questions prospects ask during a sales process, rather than chasing keywords for their own sake. What's changed is that an LLM also weighs a brand's broader footprint — consistency across LinkedIn, blog content, and citations — rather than just matching a page title to a search term.
One blog post citing the right keywords was never enough to win at SEO, and it isn't enough to win at AEO either. It takes a consistent, multi-channel presence for either engine to trust what it's surfacing.
On the content-production side, custom GPTs built around a specific brand's voice and tone can now produce genuinely on-brand copy once that foundation is set — but the input still has to start with a real point of view. Skip that step, and the output reads exactly like everyone else's, complete with the same overused em dashes and default AI tics that are quickly becoming their own tell.
Final Takeaway: Start With Your Point of View
Whether it's a coaching conversation or a piece of marketing copy, the tool only amplifies what's already there. A clear point of view, a real understanding of the person or audience, and patience with the process still come first.
AI can make a good communicator better. It can't substitute for the work of actually knowing what you're trying to say.
Recent Posts









