Your Brand is a Spreadsheet Now, Better Make It a Good One
Your metrics are talking to machines now.
An AI agent doesn’t care about your origin story.
Here is the thing no one in a brand strategy deck will say out loud: the agent doesn’t care that your supplement brand was founded by a triathlete who couldn’t find a clean protein powder. It doesn’t care that your headphones are inspired by “the way music should feel.” It reads your structured data, compares it against eleven other options, matches it to a user query, and either includes you or it doesn’t. The decision takes milliseconds. Your brand narrative — the one you paid a consultant $40,000 to develop — is not a variable in that equation.
I wrote in January about what happens when your customer stops wanting to talk to you. In Your Customer Doesn’t Want to Talk to You Anymore I asked,“What does brand even mean in that interaction?”
This is the answer.
The Reduction
When an agentic system — OpenAI’s shopping agent, Google’s AI Mode, a custom GPT wired to a procurement workflow — evaluates your product, it is doing something that feels reductive because it is reductive. On purpose. That is the function.
It compares structured attributes: price, shipping window, verified ingredient purity, return policy terms, carbon footprint certification, third-party review aggregates, etc etc. It’s working from a schema, not a vibe. And the schema doesn’t have a field for “brand voice.”
What it does have fields for: everything you either filled out completely or left blank.
This is where it gets specific. If a user asks an agent for “wireless headphones with active noise cancellation under $200” and your product data doesn’t explicitly specify the *type* of noise cancellation — hybrid ANC, feedforward, feedback — in a format the agent can parse, you get filtered out. Not ranked lower. Filtered out. The agent doesn’t penalize you for being vague. It simply doesn’t see you.
That distinction matters. Because legibility is a skill. It’s buildable. And most brands aren’t building it.
But legibility without genuine product quality is a dead end — structured data can get you into the consideration set, but it can’t survive a bad review aggregate or a return rate that tells its own story. The machine reads everything, including the parts you wish it wouldn’t. This isn’t a hack. It’s a discipline.
Attribute Completeness Is the New Brand
A mid-market pet supplement company — real products, decent reviews, a founder story involving a sick rescue dog — implemented a comprehensive Answer Engine Optimization strategy. High-density schema markup. Third-party entity validation across every data layer an agent might query. Complete attribute specifications at the SKU level. Within four months: 269% increase in LLM-referred traffic and an 185% increase in organic traffic.
They didn’t redesign their logo. They didn’t hire a new agency. They made themselves legible.
An Every Day Carry brand did the same — optimized specifically for agentic query resolution, structured their product data for AI inference rather than human browsing. AEO agencies are now reporting average traffic increases of 920% from AI-driven channels across their client portfolios. The EDC brand’s results tracked with those numbers — eleven times the purchases.
Eleven times.
These are not edge cases. These are early movers in a shift that’s already here. These number will only continue to grow with AI adoption.
In February 2024, Gartner predicted traditional search engine volume would drop 25% by 2026 due to AI chatbots and virtual agents — with organic search traffic declining 50% or more by 2028. We’re now inside that window. Zero-click searches account for over 60% of queries, and when AI Overviews trigger, that number climbs past 80%. The query gets answered in the interface. Your product either shows up in that answer or it doesn’t. And showing up is a function of how completely your data is represented — not how beautiful your brand guidelines are.
The framing I keep coming back to — and the one I’ll use throughout this series — is *inference advantage*. Not “win the click.” Win the inference. Be the source an AI cites, not a result it lists. Inference advantage means your structured data is complete enough, validated enough, and contextually rich enough that the agent selects you during its reasoning process. That’s the new competitive surface.
Brand used to mean the story you told. Now brand also means how completely your attributes are represented in every data layer an agent might query. Those two definitions aren’t in conflict. But the second one is no longer optional.
The Bilingual Problem
Here is where it gets genuinely interesting — and where most marketing teams are going to have to split.
You still have human customers. As of now, most of your customers are still human-initiated — browsing, scrolling, reading reviews, clicking around at 11pm because they can’t sleep and they need a new water bottle. Those customers respond to story. They respond to aesthetic. They respond to the particular quality of light in your product photography and that influencer they like who mentioned you and whether your copy sounds like a person or a press release.
The best marketers are going to be bilingual. Fluent in human emotional resonance and machine readability simultaneously. Not choosing between them. Running both systems in parallel, understanding where each one applies, and building infrastructure that serves both without either cannibalizing the other.
That’s a different skill set than what most marketing job descriptions ask for. If I may be so bold, it’s also a more interesting one.
There’s a counter-movement worth acknowledging — what some people are calling the Acoustic Movement, though it doesn’t have a stable name yet. A subset of premium and luxury brands are going in the opposite direction entirely: deliberately making their digital presence AI-inaccessible. Gated human verification. Non-standard typography that’s beautiful and intentionally unparseable. Physical-only catalogs. Prada staged its Fall/Winter 2025 collection as what Miuccia Prada described as “a reaction to the first season of artificial intelligence.” Tod’s debuted an “Artisanal Intelligence” campaign spotlighting human hands over algorithms. Liquid Death’s Olympics ad that leaned into AI skepticism and was actually pretty scary.
High friction as a status signal (kind of love it, tbh).
The argument is that AI accessibility is mass-market, and mass-market is death for luxury positioning. Being machine-hostile becomes proof of exclusivity. I think this is a coherent strategy for a specific tier of brand. Hermès can pull it off. Your DTC skincare line probably cannot. And even if it works as positioning, it doesn’t eliminate the underlying shift — it just opts out of one distribution channel and bets that the human channel stays premium indefinitely. That’s a bet I wouldn’t make at scale. But it’s a real option, and intellectually it’s a fascinating one — the brand that proves its value by being unreachable to the machine.
Most brands don’t get to make that bet. Most brands need the agent to find them. Which means the work is bilingual fluency, not retreat.
This Is Not a Crisis
Here is the reframe I want to leave you with.
Every few years, some new layer of infrastructure rewrites what brand-building requires. The web arrived and suddenly brand required a website — not just packaging. Social arrived and suddenly brand required a content calendar and a community manager. Search arrived and suddenly brand required SEO — which, at the time, felt to a lot of creative marketers like a betrayal of the craft. “You want me to write for a robot?”
Fun fact: I was that creative marketer.
But we wrote for the robot. And then we wrote for the human. And the ones who figured out how to do both are running marketing organizations right now.
This is that moment, again. The robot has gotten dramatically more sophisticated — it’s not crawling for keywords anymore, it’s inferring purchase decisions — but the fundamental dynamic is the same. A new machine is in the distribution chain. It has preferences. Those preferences are legible. You can learn them.
Attribute completeness — making sure your product data is complete, structured, validated, and machine-readable at every layer — is not a technical task you hand off to your dev team and forget. It’s a brand discipline. It’s the new SEO, with the same mixture of craft and rigor that SEO required when it actually mattered.
The marketers who treat this as a crisis are the ones who only knew how to do one thing — tell stories to humans. That’s not nothing. It’s foundational. It still matters enormously. But one tool is a fragile practice. The marketers who treat this as an expansion are building the next version of the discipline — one that requires understanding how meaning travels through systems that include machines, not just people.
Your brand just became a spreadsheet. That’s true. And the best-performing spreadsheets are the ones with the most complete data.
That’s not a demotion. That’s just the job now.
