Redefining Ecommerce Merchandising in the Age of AI
This article was inspired by a recent episode of “Marketing in the Age of AI,” hosted by Emanuel Rose and featuring Vinod Kumar, co-founder of Syntheum.ai.
The conversation explored how generative AI is reshaping ecommerce merchandising, why product discovery is shifting to platforms like ChatGPT and Gemini, and what merchandisers can do to stay visible in this new environment.
Merchandising Isn’t Just About the Website Anymore
For years, ecommerce merchandising meant curating a storefront - arranging categories, guiding search results, and nudging shoppers toward add-ons. It was a digital echo of traditional retail merchandising.
But here’s the catch: that center of gravity has shifted. Shoppers are no longer starting their product discovery on brand OEM websites or even Google. They’re asking questions in ChatGPT, Gemini, and Perplexity. And increasingly, they’re making buying decisions there too.
So what does that mean for merchandisers, especially at an enterprise scale?
It means the role itself is expanding - moving off-site, onto platforms controlled by large language models (LLMs). In other words, the old “optimize your site and you’re good” mindset doesn’t cut it anymore.
From SEO to GEO: A Shift Few Saw Coming
SEO once felt like the final frontier - optimize for Google, rank high, win traffic. But generative AI adoption has quietly introduced Generative Engine Optimization (GEO), a parallel strategy aimed at ensuring LLMs understand your brand and products.
As we’ve said before, optimizing for AI search and summary citation inclusion is being referred to as many things by many people: AISEO, GSO, SAO, AEO, SGE. And, believe us, the list goes on. And on. And…
Be that as it may, GEO is gaining the most popularity in 2025 glossaries and in terms of “agency speak”.
Think about it: when a shopper asks Gemini, “What are the best running shoes under $150?” the answer isn’t just drawn from Google’s index. It’s coming from how well LLMs can parse structured data, product attributes, and individual ecommerce brand visibility across the web.
That raises tough questions:
-
Does your product catalog speak clearly to machines, not just humans?
-
Are your products being compared with the right competitors?
-
And, maybe most important - what steps should you take when your brand isn’t even mentioned in those AI-driven results
Merchandising, Explained Like You’re Ten
Let’s ground this in a simple example. Imagine you search for “khakis” on a clothing site. Add an “s” at the end, and you want pants. Remove it, and suddenly it’s just a color.
Merchandising ensures that the system understands your intent - pants when you expect pants, and brown-ish shirts or jackets when you’re referencing the color.
Now take that example over to an AI platform. If you ask ChatGPT for “best khakis for summer,” how does it know whether to show pants, colors, or brands? That’s where merchandisers - and now merchandising engines - pick up the ball. They interpret intent, structured data, and connect brand offerings with human expectations.
In Vinod’s words, merchandisers are “the glue between brand and shopper.” They understand the catalog, the brand’s identity, and the shopper’s intent. With agentic ai in a supporting role, shopkeepers finally have help shouldering that burden.
Merchandisers Need Agentic Ecommerce Agents Watching Their Backs
Merchandising used to be about spreadsheets, intuition, and maybe some product data insights. But the signals have multiplied - hundreds of attributes, behaviors, and contexts. No human can reasonably connect all the dots anymore at scale.
That’s why Syntheum.ai builds agentic merchandising agents for ecommerce. Picture them as smart, well-trained assistants, not human replacements.
They handle the grunt work - processing signals, recommending pairings, flagging gaps. The human merchandiser still decides what fits the brand story, but they’re no longer buried in decision fatigue.
And here’s the subtle beauty of it: machines aren’t there to “be creative”. Their job is to free up human creativity. They do the pattern recognition so merchandisers can focus on nuance - the part that makes shopping feel personal, not robotic.
And the part that machines can’t match us humans on. Not yet, Skynet.
Closed-Loop Merchandising: On-Site Meets Off-Site
Here’s where things get even more interesting. The future isn’t just about optimizing your website. It’s about closed-loop merchandising, making sure your brand looks consistent whether a shopper is browsing your homepage or chatting with Perplexity.
That’s easier said than done. AI engines build their own context, sometimes grouping you with unexpected competitors. Vinod shared a case where a well-known footwear brand positioned itself against Nike and Adidas. But in AI-driven platforms, it was being compared with smaller sustainable brands instead. That disconnect matters - because it shapes how customers perceive you in conversations you don’t even control.
To help, Syntheum.ai has launched a free tool that lets brands assess how LLMs currently represent them. We call it “AVA” - AI Visibility Auditor. It’s like holding up a mirror: who are you compared to, how are your products categorized, and what gaps need fixing?
So What Can Brands Do Right Now?
You don’t need to wait for the future. There are tangible AI optimization steps every ecommerce brand can take today:
-
Add structured data – Product pages can’t just be pretty; they need to be machine-readable. Clear attributes, schema for ecommerce, and consistent formatting matter more than ever. It could be said, that for these reasons, schema is having its own little renaissance in 2025.
-
Experiment with LLMS.txt – Just like robots.txt for search engines, some companies are publishing guidelines for LLM crawlers. It’s not yet standardized, but it’s emerging.
-
Ask the engines directly – Unlike Google’s black box, you can literally ask ChatGPT or Gemini: “How is my brand represented? How can I improve it?” You’ll get actionable feedback - though answers vary by engine. Kick the tires on Syntheum’s own AI Visibility Auditor (AVA) for an example we’re working hard all the time to improve.
-
Balance automation with creativity – Let the machines handle scale and pattern recognition, but keep the storytelling, positioning, and creative flourishes human.
The Marketing Machine Behind the Machine
Interestingly, Syntheum doesn’t just build agentic ecommerce engines - we run our own marketing like an AI agent. Despite being a small team, we maintain a big-company presence across LinkedIn, Substack, and our own proprietary website content. The trick? AI-powered workflows.
We build custom personas trained on internal knowledge, spin one anchor asset (say a white paper or podcast appearance) into multiple formats, and push it through automated pipelines. Tools like N8N and emerging platforms like Lindy help nail the process.
It’s not just efficient - it’s philosophical. The goal is education-first marketing. Give away value through insights, build trust long before a sales call, and repurpose everything across multiple formats for maximum visibility and recognition.
Where This All This is Heading
The gosh-honest truth is: AI innovation is moving faster than AI adoption - especially in retail.
Faster than anyone can map, predict or soothsay. Standards like LLMS.txt are still experimental. Shopper habits continue shifting away from old stalwarts like Google. Ecommerce and merchandising engines are still finding their footing as assistants rather than full automation.
But the direction is clear:
-
Merchandising is no longer confined to your own site.
-
GEO is becoming as critical as SEO. Or, as many in the discipline argue, it’s simply another extension and evolution of SEO, period. And doesn’t need a unique classification at all.
-
Hybrid workflows - machines for grunt work, humans for creativity - are the sweet spot.
In three months, things may look different again. Three weeks, even. Maybe LLMs will expand checkout capabilities. Maybe new standards will emerge for structured data. Maybe another platform will emerge to challenge GPT, Claude, Gemini and Perplexity.
Merchandising With Machines, Not Against Them
It’s easy to frame AI as a threat to merchandisers, but that misses the point. By a country mile.
The real value for digital merchandisers lies in partnership. Machines handle the endless data points, freeing humans to handle the artistry.
Brands that resist this shift risk invisibility in AI-driven discovery. Those who adapt can meet shoppers wherever they are - on a homepage, in a chatbot, or in a recommendation flow built by an agent.
Merchandising isn’t dying. It’s just moving off the shelf and into the agentic, generative, conversation.
About Syntheum.ai
We help e-commerce retailers implement agentic ecommerce merchandising solutions that go beyond basic automation. By integrating truly intelligent systems into merchandising strategies, we help businesses unlock their full potential - delivering efficiencies that improve operations and redefine what’s possible in online sales.
