By 2028, an estimated 50% of all online purchases in the United States will involve an AI agent at some point in the decision process. That figure comes from industry analysts tracking the rise of agentic commerce, and if it sounds aggressive, consider that the number already sits between 25% and 30% today.
Most companies are still optimizing for a customer journey where a person searches, compares, and clicks. Increasingly, that customer journey may never happen.
ChatGPT, Perplexity, Google Gemini, and Amazon Rufus are already filtering products, comparing specifications, and completing transactions on behalf of consumers. During Cyber Week 2025, AI-influenced purchases drove $67 billion in online spending. We’re not talking about the future here. This is already happening now while most leadership teams are still debating their AI strategy.
The question this raises for any business operating across global markets is not whether AI agents will influence how customers find and choose brands. It is whether those agents can find your brand at all.
The buying journey without a human
For decades, brand visibility meant showing up where humans looked. A search result, a product page or an ad. The entire architecture of digital commerce was designed around a person navigating a screen, evaluating options, and making a choice.
Agentic commerce removes the screen entirely. When a consumer asks an AI assistant to find the best option within a set of constraints, the agent doesn’t browse. It queries product catalogs, pricing feeds, and availability data through APIs. Evaluation no longer happens visually but programmatically. If product data is unstructured, inconsistent, or incomplete, the agent may never consider that product in the first place. Visibility increasingly depends on whether machines can understand what you sell.
Cathy Hackl, CEO of Future Dynamics and Nokia’s Futurist-in-Residence, has been tracking what this means for the infrastructure that supports digital commerce. In conversation with Phrase CEO Georg Ell on the CEO-to-CEO edition of the In Other Words podcast, she noted,
The demand for connectivity at 3 a.m. is gonna be very different than demand for connectivity at 11 a.m. Those requirements are incredibly intense for the network.
The same logic applies to content. When agents are evaluating brands around the clock across every market, the content they assess needs to be available, structured, and accurate in every language at all times.
The Drum recently introduced a concept that captures this well. They call it “invocability,” a brand property that answers a simple question. Can an agent find you, trust you, and use you from anywhere without requiring a human to navigate to you first? That question did not matter before agents existed at scale. Now it does.
Salesforce’s launch of Headless 360 in April 2026 made this tangible for enterprise teams. The entire platform is now accessible through APIs and command-line tools, designed for agents to operate without a graphical interface. As Salesforce themselves advise, the platform is the data layer, the agent is the interface, and humans are increasingly the executors.
The multilingual blind spot
Here is where the conversation gets uncomfortable for global businesses. Most of the discussion around agentic commerce has focused on English-language markets. The tooling, the optimization strategies, the case studies. Almost all of it assumes a single-language environment.
AI agents evaluating brands in Germany, Japan, or Brazil are working with whatever content exists in those languages. If your multilingual content is incomplete, inconsistent, or outdated, agents may struggle to accurately represent your products in local markets. As agentic commerce expands, content quality becomes increasingly important to visibility
This matters because agentic commerce is accelerating globally. Research from Commercetools shows that over 60% of consumers in Europe are already open to AI agents helping them compare options and make purchasing decisions. Brands that cannot provide relevant content in local languages risk becoming significantly less visible as agentic commerce expands across international markets.
From robots to agents
Cathy wrote about this trajectory before the language around it existed. In 2020, she published an article in Forbes that argued brands would eventually need to market themselves to machines, not just people.
Any brand needs to think about multiple customers. They’re selling to the humans, but they’re also gonna have to sell to their agents, and then it’s gonna be agent to agent.
What makes Cathy’s perspective particularly relevant is that she saw this convergence coming from outside the traditional marketing technology world. Her background spans spatial computing, immersive platforms, and gaming, areas where the relationship between human and machine audiences has been blurring for years.
On Roblox, for example, real-time translation already allows kids in different countries to communicate in their own languages, and pricing adapts by region automatically. The infrastructure for machine-mediated, multilingual brand experiences is already being built in the places most boardrooms are not paying attention to.
Cathy also made the point that the infrastructure layer is what makes any of this possible.
Most people think infrastructure isn’t sexy. To me, it’s the sexiest thing in tech, because without that infrastructure, we can’t have the autonomous robots and the holograms and all the devices that everyone’s envisioning.
The same argument applies to content infrastructure. Without structured, consistent, multilingual content that agents can parse and evaluate, the vision of agentic commerce falls apart for any brand operating beyond a single market.
What this means for the future of global brands
The businesses that will thrive in the agentic era share a few characteristics. Their content is structured for machine readability, not just human consumption. Their product data is consistent across every market and language. And they have the content operations infrastructure to keep all of it current, accurate, and locally relevant at speed.
This is where the gap between ambition and reality shows up most clearly. Many global brands have invested heavily in their English-language digital presence while treating other markets as secondary. That approach was already costly when humans were doing the searching. When agents are doing the evaluating, it becomes a competitive liability.
But preparing for the agentic era does not mean handing everything to machines. Cathy advises,
We still need storytellers. We still need people that can work with the technical team, tell the right story, tell the right messaging, communicate in the right way.
The point is that human expertise and machine-readable content are not competing priorities. Global brands need both, working together, across every market they operate in.
As agentic commerce grows, language intelligence will become a critical part of digital commerce infrastructure. Organizations will need to ensure multilingual content is consistent, governed, and structured in ways that both people and machines can understand. The challenge is no longer simply creating content. It is making that content discoverable, usable, and trustworthy across an increasingly complex ecosystem of customers, agents, and platforms.
The cost of waiting
The transition to agent-mediated buying will not happen overnight. Human decision-making is not disappearing. But every major shift in digital commerce has changed how brands are discovered, evaluated, and selected.
The question is no longer whether AI agents will influence customer decisions. They already do.
The question is whether your brand will be visible and compelling to those agents when the moment arrives.
Every month that structured, multilingual content is live and discoverable by agents is a month where those brands are building the data foundation that agents will prefer.
Cathy Hackl was asked on In Other Words what advice she would give a CEO entering five new markets at once. Her answer was direct.
Find the right way to communicate. Don’t only rely on AI and work with experts that can help you and guide you.
That advice was always sound. In a world where your next customer might not be human, it has become urgent.
For two decades, digital marketing focused on helping customers find brands. Now brands need to help machines find them first.
Watch the full conversation
Cathy Hackl is the Godmother of the Metaverse, Global Tech Futurist, Nokia’s Futurist-in-Residence and one of Newsweek’s Top 25 AI Visionaries. On the In Other Words podcast, she shares why brands may soon need to sell to AI agents as well as humans, what a generation thinking in Robux reveals about global commerce, and who owns the digital air around you when computing leaves the screen.
