From Skepticism to Stress Test
If you’ve been following the AI search chatter lately, you’ve heard the mantra: “You need to build your brand,” and one of the top tactics is to ensure your brand is listing and promoting it everywhere.
I’ve never been against brand-building, but I’ve been wary of this recent advice to brands that it is the most crucial action to perform in AI search results. Many articles suggest that firms carpet bomb the internet, especially Reddit or other highly cited publications, for brand mentions. Too often, it sounds like a catch-all solution without proof. It echoes the same PR spin I’ve heard for decades: get your brand everywhere and the rest will follow.
But what if the user has no brand affinity or awareness? What if they’re simply asking for the best vodka for cocktails or the best CD rate available? Does “brand” really matter in those moments if they don’t know one from another? Is this simpler if there is no overarching brand exposure, you cannot be represented?
I decided to test it. And the results forced me to rethink my assumptions.
Confession: I conducted this test because my International Web Effectiveness podcast co-host, Bill Scully, baited me into it. I was ranting about my frustration with many AI Search recommendations suggesting building a brand and extensive brand promotion. I mentioned that when I tried to validate this requirement, all the sources for that activity were conveniently from branding and PR companies.
The Vodka Test — Where Brand Was Inescapable
No, we did not do this test while drinking vodka cocktails, but it might have made it more fun. It started with a simple question in multiple AI engines: “best vodka” and then “best vodka for cocktails.” I expected neutral performance attributes like smoothness, price, or alcohol content to dominate. After all, cocktails are about balance, not logos.
But that’s not what surfaced. This is a screenshot of the summary of recommendations, which includes a fairly extensive list in multiple categories based on taste tests, detailed reviews, and bartender recommendations that all referenced brands.
“At first glance, it looks like brand is the decisive factor — all the outputs are brands. But that’s a bit of a mirage. AI has to return brand names because that’s how entities are labeled. The real question is: why those brands and not others?”
“The answer isn’t simply ‘brand saturation.’ It’s whether those vodkas had computable proof — competitions, expert reviews, bartender citations — that could be used as eligibility signals. Brand is the label; the signals are the substance.”
“That’s also why Absolut failed the final cut. It wasn’t invisible — it cleared the awareness and participation gates. But it didn’t rank high enough on the validation signals. The system still returned brands, but the deciding factor wasn’t brand — it was validation.”
Instead, nearly all the recommendations were tied to brand presence in validation ecosystems:
- Competitions & Tastings: Belvedere, Grey Goose, Haku, Żubrówka — all had medals, scores, or tasting notes from reputable sources.
- Professional Endorsements: Bartenders and mixologists cited these vodkas for martinis, Bloody Marys, or citrus cocktails.
- Consumer Conversations: Forums, magazines, and social chatter reinforced the same shortlists.
- Distinctive Identity: Żubrówka (bison grass), Haku (rice), Crystal Head (agave-influenced) — brands that created liftable, memorable context.
So we asked ChatGPT for the criteria it used
One brand that surprised me was Absolut. It wasn’t missing from the datasets. In fact, it easily cleared the brand gate (high awareness, strong presence) and the participation gate (appearing in tastings and reviews).
Curious why it wasn’t in the final AI recommendation list, I asked directly: “Why isn’t Absolut included?” The answer was blunt — Absolut was present, but it didn’t pass the decisive tests for taste and blendability.
Blind tastings ranked others higher for crispness and balance. Bartenders didn’t cite it as a go-to in cocktail programs. Its profile, while solid, lacked the distinctive lift of competitors like Haku, Żubrówka, or Belvedere.
And here’s the kicker: those are not gaps you can close with marketing messaging. They live in the product, not the campaign. Which is why the AI engines filtered it out.
Your brand gets you into the game, but not onto the winner’s podium if you lack the qualifying proof.
Why Brand Shows Up in Vodka (But Not Always in CDs)
Contrast vodka with another test I’ve run: “best CD rates.”
In the CD example research I did, brand mattered far less. The AI outputs focused on hard attributes: APY, term length, deposit minimums, and insurance coverage. Wells Fargo or Chase weren’t automatically included just because of name recognition. Instead, smaller credit unions often surfaced, provided they met the eligibility gates (visibility, completeness, threshold, consensus).
So what’s the difference?
- Vodka = Superlative Subjective Query
- “Best vodka for cocktails” requires judgment. There’s no absolute metric. AI needs external validation to compute “best,” and that validation comes overwhelmingly through brands referenced in competitions, reviews, and bartender/consumer chatter.
- “Best vodka for cocktails” requires judgment. There’s no absolute metric. AI needs external validation to compute “best,” and that validation comes overwhelmingly through brands referenced in competitions, reviews, and bartender/consumer chatter.
- CDs = Superlative Objective Query
- “Best CD rate” has a clear, measurable threshold: APY. AI can compute the answer without brand affinity. The eligibility gates are structural, not brand-dependent.
- “Best CD rate” has a clear, measurable threshold: APY. AI can compute the answer without brand affinity. The eligibility gates are structural, not brand-dependent.
This shows us something critical: brand isn’t always the factor. It’s scenario-driven.
- In objective, quantifiable queries → brand is less central.
- In subjective or qualitative queries → brand presence often fills the gaps that data alone can’t answer.
What “Brand as a Gate” Really Means
Let’s be clear: when I say brand acts as a gate, I don’t mean logos or slogans. AI engines aren’t impressed by your Super Bowl ad.
What matters is brand as computable presence.
- Competitions: Have you participated in tastings, awards, or certifications that produce structured data and quantifiable outcomes?
- Reviews: Are you cited by professionals, critics, or contextually relevant and credible media?
- Conversations: Do consumers and experts mention you in forums, magazines, or cocktail guides?
- Consistency: Do these references align across sources (APY rates, tasting notes, ingredient claims)?
This isn’t hype. It’s infrastructure. Without it, AI engines can’t build consensus — and without consensus, you don’t appear.
The Case for Brand as a Gate
When AI answers a subjective query, it can’t return “smoothness” or “blendability” as the result. It has to surface a label — and that label is a brand. In this sense, brand still acts as a gate. But the deciding factor isn’t brand saturation; it’s the validation signals that sit behind the brand.
Brand as Label
Brands give AI systems something computable to recommend. Without a name, the answer can’t be packaged in a usable way.
Validation Signals Create Brand Visibility
Branding activities like entering competitions, getting reviewed, or being cited by experts generate structured data that AI engines can measure. These signals are what actually push the brand through the eligibility gates.
Risk Reduction for AI Systems
By surfacing brands tied to validation signals, AI providers can defend their output. If Belvedere disappoints, the system can point to tastings, reviews, and endorsements as evidence.
Barrier to Entry
Once brands have accumulated strong validation signals, they’re structurally advantaged. The brand name becomes a shorthand for those signals, reinforcing incumbents.
The Case Against (or at Least the Caveats)
Reframed this way, the weakness isn’t “brand itself,” but the tendency to mistake brand presence for brand validation.
Saturation Isn’t Enough
Absolut proves the point: it had awareness and presence everywhere, but without strong performance in tastings or bartender endorsements, it fell out of the final list.
Echo Chamber Risk
Over-indexing on well-known brands risks creating feedback loops where incumbents dominate simply because they’ve built more signals — not because they’re objectively “best.”
Scenario Dependency
In objective queries (like “best CD rate”), the brand-as-label doesn’t matter. AI can compute results directly from data attributes. In those cases, the validation signals are structural (APY, deposit minimums), not brand-related.
False Comfort
Believing that “brand everywhere” is enough risks diverting resources into noise. The real work is ensuring your brand name has validation signals behind it — structured data, credible endorsements, consistent references.
What Businesses Should Do About It
- Audit Your Category
Is your query landscape objective (like CD rates) or subjective (like vodka or “best laptops”)? That determines how much brand footprint matters. - Build Computable Presence
Don’t just “do PR.” Ensure your brand appears in competitions, structured reviews, and third-party validation platforms. In the CD case, ChatGPT validated consensus by using Nerd Wallet, Bankrate, and other CD rate aggregators. - Syndicate Data
Ensure your attributes (pricing, features, specs) are consistent across your site and aggregators. A brand without data completeness still fails the gates. - Test Queries Regularly
Run prompts across Google, ChatGPT, Gemini, and Perplexity. Look at who shows up and why. Are they there because of brand footprint, structured data, or both? - Invest Strategically
For subjective categories, invest in brand presence where AI looks (competitions, reviews, expert coverage). For objective categories, double down on completeness and freshness of structured data.
Closing Thought — From Vodka to Visibility
I didn’t become a “brand believer” overnight. I became a believer when I saw how AI engines struggled to answer subjective, superlative queries without leaning on brand presence as validation.
Vodka taught me that brand isn’t just storytelling. It’s the scaffolding AI uses to compute “best.”
But CDs reminded me it’s not universal. Brand is scenario-driven. In some categories, data still rules.
So don’t fall for the hype that “brand solves everything.” Instead, recognize brand for what it is: one powerful eligibility gate among many.
The businesses that thrive in AI search will be the ones that can tell the difference — and build the right kind of presence for the queries that matter most.
Brand will always be the label, but it is never the substance. AI engines don’t choose a brand because it’s famous — they choose it because the validation signals behind it are computable, consistent, and defensible.
