ShopTalk 2026 Observations: The Agentic AI Definition Gap
ShopTalk 2026 replaced "artificial intelligence" with "agentic commerce" as retail's dominant term, but the rebrand exposed a deeper problem: the industry has no shared definition of what agentic means. Two camps - answer engine visibility and internal operations automation - dominated, while brand-owned agentic product discovery on retailers' own storefronts went almost entirely unaddressed.
From "AI" to "Agentic": How ShopTalk 2026 Changed Its Vocabulary
The banners hit you before you even badged in. Agentic this. Agentic that. One vendor had it printed on their coffee sleeves. We counted at least fourteen booths where "agentic" appeared in the primary headline and the word "AI" had been scrubbed entirely, as if it never happened.
Last year was different. At ShopTalk 2025, the dominant energy was fear. People were scared. They were asking whether AI was real, whether it would replace their merchandising teams, whether the whole thing was a bubble being inflated by vendors who couldn't explain their own demos. The conversations were anxious and defensive.
This year the anxiety didn't leave. It just put on a better outfit.
"Agentic" is a more comfortable word than "artificial intelligence." It implies assistance. Partnership. A thing that works alongside you rather than instead of you. That psychological softening did real work at this conference. People were less guarded. They leaned in more during sessions. But the fundamental uncertainty about what any of this means for day-to-day retail operations? Still there. Still unresolved. The industry traded one buzzword for another and called it progress.
Two Camps, Zero Consensus: The Agentic Definition Split
Here is the thing that stood out most. Nobody agrees on what agentic means.
We spent three days pressure-testing this across keynotes, booth conversations, breakout sessions, and a Coresight Research dinner that confirmed what we'd been suspecting all day. The definition of agentic in retail has fractured into two completely separate camps, and the gap between them is enormous.
The first camp is external-facing. These are the people talking about answer engines - how a brand shows up when a consumer asks ChatGPT or Perplexity or Google AIO a product question. For this group, "agentic" means brand visibility inside LLM-driven search. It is a discovery and discoverability play. The conversation lives in the world of SEO's successor: search agent optimization, structured data, and making sure your products surface when an AI agent is doing the shopping on someone's behalf.
The second camp is internal. Ecommerce operations. Backend workflow automation. These people mean something completely different when they say agentic. They are talking about agents that execute tasks on behalf of employees - automating inventory adjustments, generating product descriptions, running pricing models, building reports. It is a productivity and efficiency play. The agent works for the retailer, not the shopper.
Both are valid applications. Neither is wrong. But there is almost zero consistency about what happens in the space between them. The applications that sit between backend operations and LLM search visibility - the actual consumer-facing, brand-owned agentic experience on a retailer's own storefront - went largely undiscussed.
That gap is not small.
The Gap Nobody Is Filling: Brand-Owned Agentic Experiences on Your Own Storefront
The biggest surprise of ShopTalk 2026 was not what people were talking about. It was what they weren't.
Almost nobody on the show floor was demonstrating or even discussing the deployment of agent-powered product discovery experiences on a brand's own website. The entire industry was either looking inward (operations) or looking outward (answer engines) and skipping over the storefront itself. The place where the brand actually controls the experience. The place where a shopper is already expressing intent with their wallet half open.
Two exceptions. Salesforce had a Herman Miller demo that attempted to show an agentic shopping experience. We walked up. We asked the sales engineer specific questions about the backend architecture - how the agent was trained, what data it pulled from, how a merchandiser would configure guardrails. Answering was a challenge. The plumbing wasn't built yet, or at least it wasn't ready to explain. But credit where it's due - they were at least trying to show what an on-site agentic experience could look like. That puts them ahead of nearly everyone else at the conference.
Sierra presented their agentic experience for Redfin - a “conversational home search” - making it real estate service-related as opposed to product-centric. While impressive, this differed from the standard fare of the conference, but at least they had a tangible example to help attendees visualize the benefits.
Shopify's presence told a different story. Their demos were focused on Sidekick - an agent designed to help merchants manage their storefronts from the backend. Building pages, adjusting settings, handling admin tasks. Useful? Sure. But it is an agent that serves the merchant, not the shopper. The consumer never sees it. The consumer never benefits from it directly. Nobody at that booth was showing an agentic product discovery experience being delivered to an actual human trying to buy something.
This is the pattern we kept running into. The industry is building agents for itself. Tools to make internal workflows faster. Tools to make brands visible to other people's AI systems. But the idea of deploying an intelligent, conversational, agent-driven experience directly on your own storefront - one that resolves shopper intent in real time, one that the merchandiser actually controls - that concept was nearly absent from the floor.
We find that strange. Not because we are biased (although obviously we are). Because it is the most logical place to start. You own the storefront. You own the data. You control the experience.
Why would you build agents for every surface except the one you actually control?
Why Cybersecurity Vendors Were Impossible to Ignore This Year
Quick detour from the agentic conversation, because something else was impossible to miss. Cybersecurity vendors were everywhere.
Not tucked into a back corner. Not sharing a booth in some overlooked hallway. They were front and center, occupying significant floor space, sponsoring sessions, and attracting crowds. This is a dramatic shift. Security used to be the plumbing of these conferences. The basement dwellers. The people you tolerated but never invited to the main stage. They were kept away from the flashy demo areas because nobody wanted to talk about data governance when they could watch a hologram recommend sneakers.
That changed this year. And it changed because of the very agentic systems everyone else was hyping. You cannot deploy autonomous agents - agents that access product catalogs, customer profiles, behavioral data, inventory systems, pricing engines - without locked-down data pipelines. You cannot hand an AI system the keys to your commerce infrastructure and then shrug at consent management. The data requirements of agentic commerce dragged security out of the basement and onto the main floor. About time.
Syntheum’s own Vinod Kumar, CEO, and Lindsay Vaughn-Luma, Head of Customer & Partner Success
Who Actually Owns the Term "Agentic Merchandising" and Does It Matter?
We should talk about the phrase agentic merchandising for a moment.
It is getting crowded. We noticed competitors at ShopTalk beginning to claim the term. New decks. New landing pages. New positioning statements that sound familiar if you have been paying attention to this space for more than six months.
Our choice of definition helped bring “agentic merchandising” into the ecommerce zeitgeist before the industry was paying attention to the concept, let alone fighting over the terminology.
While most of these companies were still figuring out how to rebrand their AI messaging, we were deploying no-code, merchant-controlled agentic experiences on live storefronts for major retailers like YETI. Not developer-driven front-end projects that take months to scope and ship. Not point solutions bolted onto existing search tools.
Syntheum builds foundational operational infrastructure designed to let merchandising teams iterate faster, at lower cost, with full control over an optimized experience. That is not a complaint about the competition catching up to the language. It is a statement about how long we’ve been doing the work.
The irony of the show floor is worth noting, though. Despite the sudden popularity of the language, there was very little actual presence around what agentic merchandising looks like in practice. The words were everywhere. The working demos were not. The booths that used the phrase could not, in most cases, walk you through a live implementation or explain how a merchandiser would interact with the system day to day.
McKinsey's numbers are pretty damning here - 71% of merchants say AI merchandising tools have done little to nothing for their business, and 61% admit their org just isn't ready to scale it.
It was positioning without proof. And McKinsey's research backs this up - 71% of merchants say AI merchandising tools have had limited to no effect on their business, and 61% say their organization is not prepared to scale AI across merchandising.
Frequently Asked Questions
1. What is agentic commerce?
Agentic commerce is a broad term describing the use of AI agents in retail and e-commerce. At ShopTalk 2026, it was used to describe everything from answer engine visibility and LLM-based product discovery to internal backend workflow automation. The term has no single agreed-upon definition across the industry, which is part of the problem.
2. What is the difference between agentic commerce and agentic merchandising?
Agentic commerce is the umbrella category. Agentic merchandising is more specific - it refers to using AI agents to handle product ranking, search optimization, category management, and intent-based product discovery. Syntheum has been publishing under the term "agentic merchandising" for over two years, predating most competitors' adoption of the phrase.
3. Why aren't more retailers building agentic experiences on their own storefronts?
Most vendors and platforms are focused on either internal operations (making retailers more efficient) or external visibility (making brands discoverable in AI answer engines like ChatGPT and Perplexity). The storefront itself - where the brand controls the experience and the shopper is actively expressing intent - has been largely overlooked. Infrastructure gaps, unclear ROI models, and the novelty of the technology all contribute.
4. What role does cybersecurity play in agentic commerce?
Autonomous agents require access to product catalogs, customer profiles, behavioral data, pricing engines, and inventory systems. That level of data access demands locked-down pipelines, consent management, and governance frameworks. Cybersecurity's rising presence at ShopTalk 2026 reflected this reality - you cannot deploy agents at scale without solving data security first.
5. How is agentic AI different from traditional AI in retail?
Traditional AI in retail has largely been reactive and tool-based - recommendation engines, rules-based search ranking, automated reports. Agentic AI operates with greater autonomy, executing multi-step workflows, making decisions within merchandiser-defined guardrails, and adapting in real time to shopper behavior and demand signals. The shift is from AI as a feature to AI as an operational layer.
What ShopTalk 2026 Proved (and What It Didn't)
Retail changed its vocabulary at ShopTalk 2026. It did not change its execution. Calling something agentic instead of AI does not solve the integration problems that merchandising teams face every day. And until more of this industry starts building for the consumer standing on the storefront - not just the agent sitting in the backend - the distance between booth marketing and deployed reality is going to keep growing.
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.
Empower Merchants with Ease and Intelligence
Syntheum is the Semantic Merchandising Platform for Agentic Commerce - powering onsite search, conversational shopping, and AI discovery through one merchandising brain your team controls.