GEO went from nonexistent to the #2 B2B marketing investment priority in twelve months, but the gap between the spending and the ability to measure it should worry anyone writing the check.
Twelve months ago, “Generative Engine Optimization” didn’t show up as a budget category in Wynter’s annual CMO buying survey. Zero percent. Not because CMOs were ignoring it. Because the category didn’t exist yet.
In the 2026 survey, GEO landed at #2 on the investment priority list. Thirty-four percent of B2B SaaS CMOs now allocate budget specifically toward making their brand visible in AI-generated answers. Second only to AI/automation tools themselves.
A category going from nonexistent to second place in a single year is not normal, even in an industry that loves chasing new channels. And the speed alone should raise a question that very few people selling GEO services want you to ask: what exactly are these CMOs buying?
The fear is real, even if the playbook isn’t
The underlying anxiety makes perfect sense. Wynter’s data shows 84% of CMOs now use AI tools for vendor discovery, and 68% start their searches with AI before they ever touch Google. If your brand doesn’t appear when a buyer asks ChatGPT or Perplexity for recommendations in your category, you’re invisible at the moment that matters most. The report puts it bluntly: if the AI doesn’t know you, neither do 84% of your buyers.
That’s a real problem. The panic is justified.
But justified panic and sound strategy are different things. And right now, the GEO market is long on panic and short on proof.
Most of what’s sold as GEO is just digital PR with a new label
Look at what GEO agencies and consultants actually deliver. The typical package includes: getting mentioned in roundup articles on authoritative sites, placing expert quotes in industry publications, building backlinks from high-domain-authority sources, and generating reviews on platforms like G2 and Capterra.
That’s digital PR. It was digital PR last year. It was digital PR five years ago. The tactics haven’t changed. What changed is the pitch deck. Where agencies used to say “this will improve your domain authority and search rankings,” they now say “this will make LLMs recommend you.” Same deliverable, higher price, newer acronym.
That’s not to say the tactics are worthless. Getting mentioned frequently across authoritative sources probably does influence what LLMs surface. The training data has to come from somewhere. But the causal chain between “we placed three articles on MarTech Today” and “ChatGPT now recommends us for ABM platforms” is long, murky, and largely unmeasured.
The attribution problem is not a small detail
Here’s where the conversation gets uncomfortable. Traditional SEO has clear, if imperfect, attribution. You can track rankings, clicks, conversions. You can see which pages drive traffic and which queries bring buyers. Paid search is even cleaner. You spend money, you see results, you calculate ROI.
GEO has none of this. There is no Google Search Console for ChatGPT. There is no rank tracker for Perplexity. When a CMO asks Claude to recommend marketing automation platforms and your name comes up, nobody on your marketing team knows it happened. Nobody can attribute the downstream demo request to that AI interaction.
Some companies are trying to build measurement proxies. They run periodic queries against major LLMs and track whether their brand appears. They survey new customers about how they discovered the product. They look for correlations between PR placement and LLM mentions. These are reasonable efforts. They’re also nowhere close to the attribution rigor that would justify the budget reallocation Wynter’s data describes.
Thirty-four percent of CMOs are making meaningful investments in a channel they cannot reliably measure. That’s not strategy. That’s a hedge against fear.
What the few serious operators have figured out
There are companies approaching GEO with actual discipline, and what they’re doing looks different from the standard agency playbook.
The first thing they’ve recognized is that LLMs don’t work like search engines. Google ranks pages. LLMs synthesize information across sources and produce a single answer. Being “optimized” for an LLM isn’t about ranking for a keyword. It’s about being the brand that appears consistently across enough authoritative sources that the model can’t give a coherent answer to a category question without mentioning you.
That sounds like brand awareness, and it partly is. But the specifics matter. The companies showing up reliably in AI answers tend to share a few traits: strong review presence on structured platforms (G2, Capterra, TrustRadius), frequent mention in editorial content that isn’t sponsored or paid, a clear and distinctive point of view that gives writers a reason to cite them, and robust technical documentation that LLMs can parse cleanly.
Notice what’s missing from that list. Gated content. Sponsored webinars. Display ads. None of the standard demand gen machinery contributes to LLM visibility. The inputs that matter are the ones that look least like marketing and most like earned reputation.
The real question behind the budget line
The Wynter data creates a useful pressure test for any CMO currently investing in GEO. Not “should we do this” but “do we know why we’re doing this, and would we keep doing it if the fear subsided?”
If the answer is that you’re building genuine authority, earning editorial mentions, strengthening your review presence, and creating content clear enough for both humans and machines to parse, then the investment makes sense regardless of what you call it. Those are durable marketing activities. They were valuable before GEO had a name.
If the answer is that you’re paying an agency to “optimize for AI” because you’re terrified of being left behind, and you have no way to measure whether it’s working, that’s a different situation. That’s a budget line driven by FOMO, dressed up in a category label that’s barely a year old.
The 34% of CMOs investing in GEO are probably right that AI visibility matters. They’re probably wrong about how much of what they’re buying actually delivers it.