AI Agents Are Beginning to Buy on Behalf of Users
“Within 12 to 18 months, your customer will no longer visit your website,” warns Paolo M. Luino. An ever-growing share of online purchases is now being managed directly by AI agents. “The website no longer needs to convince a person, but an algorithm.”
According to Paolo M. Luino, General Manager EMEA at Kitsune and a leading international expert in digital strategy, the new goal for websites is not to be visited, but to be selected. After anticipating the crisis of traditional SEO and the shift towards Search Ecosystem Optimization, Luino now identifies the next major evolution: the transition from visibility to transaction.
The Shift in Decision-Making
“First, the search engine changed. Then, the ecosystem changed. Today, the entity making the decisions is changing: it is no longer the person, but their AI agent,” Luino explains.
This shift is already visible in the platforms millions of people use every day. AI tools are no longer limiting themselves to answering questions; they are starting to take action. A user can now ask for product recommendations, delegate the comparison of alternatives, get selections based on precise criteria, and ultimately reach a final choice without ever opening a website. It is no longer about browsing; it is about delegation.
“The user expresses a need in natural language, and the agent does the rest,” says Luino. “It analyzes the web, compares offers, checks availability, cross-references reviews, and delivers a pre-filtered decision. The customer intervenes only at the end, when the process has already been completed.”
This marks a definitive breaking point. In previous transitions—such as from desktop to mobile—the tool changed, but human behavior remained central. Today, however, the decision-maker itself is changing. “This is the first transition where the choice isn’t constructed by a person, but by a system that reads data, interprets it, and compares it in seconds.”
From Destination to Queryable Source
As a result, the corporate website stops being a destination and becomes a source to be queried. Often, without ever receiving a human visitor. In this new landscape, everything designed to attract, engage, or persuade a human user loses much of its effectiveness.
“An agent doesn’t get persuaded,” Luino points out. “It doesn’t look at design, it doesn’t read emotional headlines, and it doesn’t dwell on storytelling. It seeks information. And it wants to find it immediately, in a clear, structured, and comparable format.”
5 Actionable Interventions to Redesign Your Architecture
This requires a complete rethink of website architecture. Moving away from a digital storefront toward a queryable information system. To achieve this, Luino suggests five concrete, immediately applicable interventions:
- 1. Structured Data: Everything a company offers must be in a structured, machine-readable format. This means utilizing Schema markup, organizing information, and producing clear, unambiguous content. “If data isn’t structured, for an agent, it’s as if it doesn’t exist.”
- 2. Absolute Clarity: Every page must answer a few essential questions without hesitation: what you offer, to whom, at what price, and how. Eliminate implicit steps and distracting storytelling. “An agent is extremely fast, but it has no patience: if it doesn’t find the answer immediately, it moves on.”
- 3. Technical Queryability: Product availability, terms, timeframes, and specifications must be accessible and verifiable. This requires updated feeds, queryable systems, and real-time information. “The shift is from a company that gets read to a company that gets queried.”
- 4. Distributed Reputation: AI agents don’t stop at your website; they cross-reference external signals. Reviews, citations, and mentions on authoritative sources become crucial for establishing trust. “If the web isn’t talking about you, the AI won’t trust you.”
- 5. Information Freshness: Maintaining updated content, recent data, and clear time signals is vital. Agents reward what is verifiable over time. “Outdated content is content that loses value in the eyes of the AI.”
The Power of Consistency
Added to these elements is a strategic factor: consistency. The clearer a company is in defining its scope, the more it is recognized as a benchmark in that category. AI agents learn by association, and when that association solidifies, a cumulative mechanism kicks in, reinforcing those same choices over time.
“It’s a flywheel,” explains Luino. “The more you are selected, the more you become the natural choice. And the more you become the natural choice, the more you continue to be selected.”
Adapt or Be Left Behind
In recent months, Luino summarized the shift with a clear formula: the challenge is no longer getting found, but becoming the source. Today, that transition is evolving even further. “It’s no longer enough to be cited: you have to be chosen,” he concludes. “Because very soon, people will no longer be visiting sites and comparing offers. AI agents will do it for them. And at that point, you are either inside their decision-making process, or you are out of the market.”
For businesses, the time to adapt is limited. It isn’t about introducing new tools, but rethinking existing assets so they are comprehensible, queryable, and trustworthy for an automated system. “Those who continue to design their site for a human user are working for a world that is disappearing,” warns Luino. “The new goal is not to be visited, but to be selected.”