Hi everyone, this is Neo.

Quick question before we start: if your boss asked you right now, “we’ve put a lot of money into AI — is it actually working?” would you have hard numbers to put on the table?

Probably not. But here’s the thing — that’s not just a you problem. The top marketing executives at America’s biggest brands can’t answer it either.

On August 5, TransUnion — one of the big three credit bureaus, a company that literally lives on data — released a study of 100 senior marketing and technology leaders at major US brands, run through UTA’s brand advisory division. SEJ’s Greg Jarboe took the data, interviewed TransUnion’s Matt Spiegel (EVP of TruAudience Growth Strategy), and landed on the real story: marketers have never been more confident about AI, and they’ve never been less able to prove it works. TransUnion calls it the “confidence-readiness paradox.”

Today I’m breaking down the study’s core numbers, then giving you three things you can do this quarter: how to use this data to win GEO budget from your boss, why “fix your data before you buy another tool” applies to independent sites just as much as enterprise brands, and how to build incrementality testing into your AI search measurement.

1. First, let’s be clear about what this study actually is

So you can verify it yourself, here’s the provenance:

• Who ran it: TransUnion (NYSE: TRU), fielded by UTA Advisory, the brand advisory arm of United Talent Agency
• Who was surveyed: 100 senior marketing and technology leaders at major US brands, director level and above
• Released: August 5, 2026 (Chicago)
• Who decoded it: Greg Jarboe of SEJ, in conversation with Matt Spiegel, EVP of TruAudience Growth Strategy at TransUnion

One caveat on the sample: as PPC Land noted in its coverage, with 100 respondents each percentage point represents roughly one person, so treat these figures as directional signals, not precise population estimates. But the direction is unmistakable — and it lines up almost exactly with what other AI measurement studies have been finding for months. That cross-validation is worth more than any single number.

2. The core finding: what the “confidence-readiness paradox” actually looks like

Here are the numbers that matter, in one table:

Dimension	Data
Expect AI-enabled marketing investment to grow over the next 12–24 months	89%
Confident they’ll hit their AI goals	64%
Rate their people/talent readiness as high	only 42%
Rate their data and process readiness as high	only 36%
Have enough visibility into platform-level AI to make optimization decisions with confidence	only 48%
See meaningful ROI from AI-driven marketing	only 53%
Have fully scaled AI across the enterprise	only 13%

See the gap? Confidence (89%, 64%) versus readiness (42%, 36%) — there’s a canyon in between. And “only 53% see meaningful ROI” is the line PPC Land used as its headline — the most uncomfortable sentence in the whole study.

The breakdown of where AI is actually being used makes the picture even sharper:

• Adoption is huge at the execution layer: 92% use AI for productivity, 89% for basic content and creative development, 83% for creative optimization
• But more than half of companies haven’t touched the higher-value decisioning and performance use cases
• 75% say AI reduced manual effort — AI reliably saves time; it only occasionally drives results

One sentence summary: in marketing today, AI’s best job is being the intern, not the strategist.

3. Why the ROI isn’t there: data is the first wall you hit

When asked to rank the three gaps — people, process, data — Spiegel’s answer was refreshingly blunt:

“If I had to rank them, I’d put data first. AI can compensate for a lot of things, but it can’t compensate for incomplete or disconnected data. If you’re feeding AI incomplete information, you’re going to get outcomes that are less predictive than you hoped.”

The data backs him up completely:

• 67% report siloed or fragmented data systems
• 42% have incomplete or missing data
• 31% struggle with data latency
• 69% say walled-garden blind spots limit their ability to evaluate AI’s effectiveness
• 70% say cross-channel blind spots make it hard to track AI’s impact across the customer journey

And the way companies measure AI makes it worse:

How AI marketing impact is measured	Share
Estimated cost/time savings (hours saved)	65%
Marketing mix modeling (MMM)	42%
Mostly stakeholder perception and anecdotal feedback	22%
Not measured at all	14%

In plain English: most companies measure AI by “how many hours we saved,” not “how much money we made.” When you make investment decisions off a cost-savings metric, you stay stuck at the time-saving layer forever. It’s a closed loop: deploy AI to save time → measure it by time saved → keep funding the same logic → never touch real growth value.

4. So what does this have to do with SEO and GEO?

This is the most important section of the article — and the most valuable part of Jarboe’s analysis.

Every AI citation study and GEO report out right now (including the ones I’ve written about) is essentially reverse-engineering a system that was never designed to be measured from the outside. How Google’s AI Overviews, ChatGPT, or Gemini actually choose their sources is something we can only infer — through scraping, samples, and guesswork.

What the TransUnion study hands us is independent confirmation from outsiders: 100 marketing leaders with zero stake in SEO’s internal debates are describing the same problem — the whole industry has a measurement blind spot.

Spiegel put it directly: “AI search is another example of why independent measurement is becoming more important.” Consumers increasingly discover brands through AI-generated answers instead of traditional search — and that entire journey sits inside the blind spot of every legacy attribution tool.

What this means for us: AI search visibility isn’t an isolated SEO sub-discipline. It’s one instance of a company-wide measurement failure, showing up in the search channel. CMOs are already naming this problem in board meetings. The SEO team that frames GEO as the search-side fix for a measurement gap the CMO already knows exists gets the budget sign-off. The team that pitches it as a shiny new discipline gets a polite “we’ll circle back.”

5. Three things you can actually do this quarter

Move one: stop pitching AI search visibility as a standalone budget line.

Walk into the next budget conversation with the TransUnion numbers: 89% of brand leaders are scaling AI, but only 53% can prove ROI and 70% admit cross-channel blind spots. Then say: “AI citation tracking is the search-specific piece of the measurement gap you already know you have.” One sentence changes who signs the check — you’re no longer the vendor asking for money for a new concept, you’re the fix for a hole they already know exists.

Move two: audit whether your data is connected before you buy another AI visibility tool.

Spiegel’s advice to a team with a fresh AI budget: spend the first dollar on data quality and identity resolution, not another application. The same logic applies to an independent site — if GA4, Search Console, your CRM, and your ad accounts don’t talk to each other, a fancy GEO tool will report activity, not impact. Garbage in, dashboard out.

Move three: build incrementality testing into your AI search measurement now.

Right now only a minority of marketers use marketing mix modeling or incrementality testing for AI, mostly because reporting time saved is easier and faster. But when these standards become the norm, the people who ran the tests first are the ones defining the standard. If your SEO team can show “this incremental traffic and conversion lift came specifically from AI search,” that number becomes your strongest card in next year’s budget negotiation.

And Spiegel’s closing line — which I think deserves the headline more than anything in TransUnion’s press release:

The biggest mistake he sees isn’t a bad AI strategy. It’s leaders asking “what’s our AI strategy” instead of “what business problem are we solving.”

Translated into SEO terms: the teams still asking “how do we win more AI citations” should first ask “what business problem are we solving.” Teams that answer that question differently end up doing completely different GEO.

Neo’s take

First, the most valuable thing this study gives Chinese independent-site sellers is negotiating language. You don’t need a 100-executive sample — you need numbers like “89% vs 53% vs 36%” that a boss or a client can absorb in five seconds. Data is the universal currency when you’re talking to overseas CMOs or your own partners. Citing a US-brand executive survey to justify your GEO budget beats saying “AI search is important” a hundred times.

Second, the “data before tools” ordering hits Chinese sellers especially hard. Most independent sites have GA4 and Search Console “connected” in the loosest sense — forget about the CRM and ad data. Spiegel’s right: AI is a force multiplier, not a shortcut. Fragmented data in, unreliable AI output out. Spend the first dollar on data quality — I can’t endorse that advice strongly enough.

Third, the “time-saved vs money-made” insight is worth taking back to your own weekly report. If you’re still reporting “saved X hours” to your boss, switch this week to “drove X incremental traffic/conversions.” How you measure determines where resources flow — that’s not office politics, it’s a pattern this data keeps confirming.

Fourth, on where GEO fits: I agree with Jarboe — don’t treat GEO as a new continent; it’s the search-channel projection of a measurement gap. For Chinese sellers this means GEO budget shouldn’t come out of the “AI is new so we should spend on it” bucket. It should come out of the “our attribution is breaking down and we need to fix search-side measurement” bucket. To a boss, one is a cost-center stop-loss; the other is a novelty experiment. Same money, completely different fate.