We’ve now run Audit*E reports for dozens of organizations, from tech startups to healthcare systems to financial services firms. And there’s a pattern to how these conversations unfold.
The marketing leader comes in expecting to discuss brand messaging and content strategy. But they leave realizing they’ve been playing an entirely different game than they thought.
The Schema Confusion
Almost every initial presentation includes some version of this exchange:
- Marketer: ”I’m trying to understand why organizational structure matters to LLMs. I mean, we want to tell our story through our narrative.”
- Us: ”This is about the back-end code. Schema markup helps AI understand what it’s looking at.”
- Marketer: ”But won’t that affect the user experience? I need people on mobile to get to information quickly.”
- Us: ”Consumers never see schema. It’s exclusively for machines.”
- Marketer: ”Oh. Oh. Okay, that changes everything.”
This happens in nearly every first meeting. Why? Because marketers have spent careers thinking about front-end user experience. Generative engine optimization (GEO) requires them to also think about back-end machine comprehension. The code that most people never see determines whether AI platforms can understand your organization well enough to recommend it.
The Citation Shock
You may wonder: How do LLMs learn about brands? How do I optimize my brand for ChatGPT?
When we show clients which external sources AI platforms most frequently cite when discussing their brand, the response is almost always some version of: “That’s… deeply strange.”
We’ve seen:
- Real estate listings cited more than official partnerships
- Individual employee LinkedIn posts outranking corporate communications
- Industry directories that haven’t been updated in three years treated as authoritative
- Former vendors cited more frequently than current ones
One client discovered AI was citing a conference presentation from 2019 more than any of its 2025 content. Another found a critical review on a niche forum was the most-cited source about their customer service.
The sources AI trusts are rarely the sources marketers are investing in.
The Multi-Account Chaos
Here’s a scenario we’ve encountered at multiple organizations: Different departments or locations have created their own Google Business Profiles. Marketing coordinators have set up separate accounts. Individual practitioners have their own listings. Nobody has a master view of what exists or who controls it.
One marketer described it perfectly: “I’ve been picking up rocks with so many spiders underneath. There should be ONE account with everything managed centrally. Instead, I’m discovering accounts I didn’t know existed set up by people who no longer work here.”
The result? Inconsistent naming conventions. Contradictory information. AI platforms seeing multiple conflicting signals about the same organization. AI platforms using wrong sources about brand.
Traditional marketing could survive this kind of fragmentation. AI optimization cannot.
The Platform Blind Spots
We regularly surface platforms that marketers have completely ignored, but AI happens to trust deeply.
Reddit comes up often. Most marketers are nervous about it: “There’s misinformation. The user-generated content is unpredictable. The stuff that gets picked up is the juicy, controversial stuff.” But Reddit is one of the top sources feeding major LLMs and driving what many businesses view as “next generation SEO.”
Ignoring it doesn’t make it go away. It just means you’re absent from the conversation.
Bing is another common blind spot. With Microsoft’s Copilot integration, Bing business listings suddenly matter. But we regularly find organizations with robust Google presence and zero Bing presence. One client had 15 well-maintained Google Business Profiles. Bing listings? None.
The Domain Authority Recalibration
When we show domain authority scores, marketers often react like they’re seeing a bad report card. “That’s an F” is a common first response to a score of 35/100.
Then we have to recalibrate expectations: “A score of 100 is Wikipedia or Harvard. You’re competing in commercial space where 40-50 is strong. You’re not failing. You’re just not at the ceiling.”
Traditional marketing metrics trained people to expect scores in the 80s and 90s. Domain authority works differently, and most marketers don’t know the benchmarks.
The “Who Controls This?” Problem
Almost every brand AI analysis audit reveals ownership ambiguity. We hear things such as:
- “I think IT owns some of our listings, but I’m not sure which ones.”
- “Our web developer might have access, but they’re a contractor.”
- “Different locations manage their own profiles.”
- “I know we have a Healthgrades presence, but I don’t know who set it up.”
One marketer summed it up: “I feel like I need to translate all of this for the web people, their vendors, our IT department, and I don’t really know what I’m telling them.”
GEO requires coordination across teams that traditionally haven’t worked together. And it requires someone to actually own the AI visibility strategy.
The Weight Question
This question comes up in nearly every initial presentation: “If we’re going to focus on two or three things to see a lift, where should we focus our energy? What has the most weight?”
It’s a reasonable question. Marketers are used to prioritizing tactics based on expected ROI.
But GEO doesn’t work like traditional channel marketing. It’s infrastructure, not campaign work. Schema markup matters. Citation consistency matters. Listing accuracy matters. Technical site performance matters. There’s no single “high-impact” tactic. Rather, it requires a systematic cleanup of foundational issues most organizations didn’t know they had.
What Changes After the First Report
By the end of these initial presentations, we can see the mental shift happening:
- Before: ”We need better content and messaging for AI platforms.”
- After: ”We need to audit our entire digital infrastructure to understand what signals we’re actually sending.”
- Before: ”Can you help us optimize for ChatGPT?”
- After: ”Can you join our web development meetings to make sure they’re implementing this correctly?”
- Before: ”This is a marketing project.”
- After: ”This requires coordination across marketing, IT, web development and every department that touches our online presence.”
The Real Complexity
GEO isn’t complex because the tactics are sophisticated. It’s complex because it requires:
- Cross-functional coordination most organizations aren’t structured for
- Technical infrastructure work that marketing teams don’t traditionally own
- Citation management across platforms they may have never thought about
- Consistency at scale when different departments have been operating independently
- Understanding machine logic in addition to human psychology
One marketer captured it perfectly: “I don’t speak developer. I kind of know what I know, and I know there’s stuff missing, but I don’t necessarily have the language to relay that. But I’ll know when I see if it’s not working.”
That’s where most marketing leaders are right now with GEO. They know it matters. They know it’s different from traditional SEO. They know they need a GEO assessment and do things differently.
But they’re discovering that the complexity isn’t in the strategy. It’s in the execution across systems and teams they don’t fully control.
The Question Nobody Asks (But Should)
Not once has a client asked: “What is AI saying about our competitors?”
Everyone wants to know what AI says about them. Almost nobody asks what AI is learning about the competitive landscape. But that’s the real opportunity.
While your competitors are also figuring this out, understanding how AI platforms are currently representing your entire category gives you a roadmap for differentiation.
The organizations that figure out GEO can shape how AI understands their entire industry.
Bospar’s Audit*E tool analyzes how AI platforms represent brands across ChatGPT, Perplexity, Gemini and other LLMs. Wondering what AI actually knows about your organization and what it’s learning from sources you didn’t even know existed? That’s exactly what Audit*E reveals.
