Your Customer's Next Visit Might Not Be Human

Every visit now passes through an AI filter

Rokas Stankevičius, founder of AIclicks.io, puts it bluntly: every interaction a customer has with the internet is moving behind an AI filter. Phones, feeds, smart speakers and car dashboards now summarize and recommend products before a shopper ever sees a brand name. At the same time, the line between Google and ChatGPT is dissolving — Google's AI Overviews and AI Mode are turning search itself into an answer engine.

His EcomExpo talk walks through a case where a single ChatGPT recommendation, engineered through content and product data, turned into a $100,000 deal for one e-commerce brand.

The opening for smaller players: GEO on top of SEO

Chris Kilin, co-founder of Glara, saw the same bottleneck forming while scaling e-commerce brands at Razor. His argument is that AI-driven discovery actually opens a door for smaller players: assistants pre-filter products on relevance and product truth, not ad budget.

But it breaks the old measurement toolkit — clicks, last-touch attribution and rankings explain less and less about why a brand was or wasn't recommended. His prescription: GEO on top of SEO. Clean catalog structure, consistent attributes, accurate pricing and availability, clear shipping and returns — product information machines can trust and compare.

The great traffic collapse

Ryan Robinson, whose content reaches 500,000 readers monthly, calls it the great traffic collapse. In his framing, discovery is no longer a funnel but a constellation: a TikTok comment thread, a Reddit recommendation, a ChatGPT comparison, and only then maybe a Google search to confirm the brand exists.

His conclusion is uncomfortable for anyone still reporting organic sessions to the board: rankings and traffic are now lagging indicators. Citations, branded queries and conversational share-of-voice are what predict revenue.

The durable moat is being a brand people search for by name — everything else is a rented audience.

The contrarian: signal consistency, not shiny objects

Not everyone on the lineup thinks the answer is chasing the new shiny thing. Kaspar Szymanski, a former senior member of the Google Search team, insists 2026 is evolution, not revolution.

Signal consistency. Anything else is a distraction.

Large platforms fail, in his experience, not because they missed a trend but because their signals contradict each other. The brands that win are the ones with a disciplined, data-driven decision process — the ones merely chasing buzzwords fail.

The autopsy: AI exposes broken processes

And then there's the human layer. Balázs Horváth of VisualLabs spent twelve months turning his own company into an AI-operated business and is bringing the autopsy to Vilnius: 17 dead agents, 20+ missed deadlines, tens of thousands of euros in wasted licenses.

AI doesn't fix broken processes, it exposes them.

His core lesson applies directly to commerce: if a new hire can't run your returns process from documentation alone, no AI agent will either. His counter-intuitive prediction: mid-sized brands with disciplined data will outperform large brands carrying decades of integration debt.

The trust problem: used, not just tolerated

Emilija Antanavičiūtė, behavioral researcher at ISM, closes the loop from the customer's side. Businesses scrambled to ship AI chatbots as both a branding move and a cost cut — and users noticed.

Her session digs into the research on why people distrust unfamiliar technology (the "black box" problem, anthropomorphism) and how to design AI features customers actually use instead of tolerate.

The through-line: it's an audit, not a feature

The through-line across all six: agentic commerce isn't a feature you bolt on. It's an audit of everything underneath — your product data, your processes, your measurement, and your customers' trust.