Shopping trends follow technology trends. Hot on the heels of each tech milestone of the digital age—the internet, social media, mobile—came marketers. Titans like Amazon, eBay and Temu were built by leveraging new opportunities to reach customers.
The wave of transformation that will be driven by the widespread adoption of AI agents could potentially be even more dramatic.
Agents are best thought of as the next leap forward in generative AI. Rather than simply thinking and talking like ChatGPT, agents are capable of taking action and performing complex tasks.
According to a recent research paper, Are AI Agents Interacting With Online Ads?, AI agents are already looking at advertising to decide whether it could help them answer questions.
The paper also reports that the number of browser-based searches could drop by 25% by the end of this year in favor of searches carried out on our behalf by agentic or generative applications.
So, what steps are businesses taking to prepare for this opportunity? And what does anyone responsible for marketing a business, from corporate marketers to small business owners, need to do to make sure they’re ready?
How Agents Shop And Make Buying Decisions
When explaining what agents can do, one of the most frequently cited examples is their ability to make buying decisions. By using a browser or interfacing APIs, they can, in theory, vastly speed up the process of making a purchase. While some of us might enjoy shopping, going through numerous sites to compare prices, shipping times, or return policies is still a time-consuming activity for humans.
Agents are well suited to this because, with their analytical and language capabilities, they can quickly create an overview of our options, along with the pros and cons of choices, individually tailored to what they know about our needs.
AI agents already exist that can help users by creating recipes to cook and shopping for ingredients, as well as scanning multiple sites looking for and booking the best travel deals, have already been demonstrated.
The research examined behaviors of popular AI tools with agentic capabilities, including OpenAI Operator and open-source models.
Among the findings was a propensity for agents to focus on structured data, such as lists of statistics, for the content they analyzed. It also discovered that agents using computer vision capabilities to “see” websites like us are easily tricked into clicking pop-ups.
You can read a more detailed technical breakdown of the findings here. And the full impact of these changes is likely to take some time to take effect, with AI agents possibly acting very differently tomorrow than they do today.
However, in terms of considering how this will impact everyone trying to connect with customers or sell online, some clues are already starting to emerge.
So, will the next Amazon, eBay or TikTok Shop be a business that beats all competition when it comes to capitalizing on these behavioral changes? Let’s look at some of the ways this could happen.
How Can Businesses Leap On This Opportunity?
For marketers, this might seem to turn one of the fundamental principles of digital marketing on its head.
Rather than persuading humans, the priority becomes persuading machines that are advising humans. The goal is to convince them that our ideas, messages, products or services are the best response to their users’ requests.
The first clue we have is that we should focus on structured, machine-readable information. This might be because machines know we like information backed by facts and stats. But despite their ability to interpret complex expressive human languages, they still instinctively look for information in their native language of numbers.
The paper also suggests (although doesn’t definitively state) that well-sourced information with quality citations is more likely to attract the attention of agents.
And it’s thought that agents are also more likely to be attracted to information that appears to be direct answers to questions, written in a clear and well-structured way.
On the technical side, it would be a good idea to see if APIs and feeds can (and should) be made agent-friendly, opening up streams of data on products, prices or availability. This might involve marketers working with digital or e-commerce teams to make sure everyone’s on the same page.
And this cross-organization cooperation will also be needed when it comes to understanding how agents are interacting with and accessing your data. Developing tools for detecting and tracking the difference in human and machine behavior would be a great way to get a head start today.
Of course, it would be premature to say that “traditional” (not optimized for agents) SEO is dead or even dying. After all, web browsing is expected to be one of the most important tools in an agent’s kit. So/ they will get served up content that has been well-optimized first, just like the rest of us. For now, at least.
But there will be other ways of attracting the attention of agentic AI; exploring and preferably mastering these should be a priority for anyone involved in marketing right now.
And After That?
It’s very hard to predict how any hugely transformative technology will work out long-term, and agentic AI is no different. When marketers started SEOing content decades ago, it was largely a technical exercise to count keywords. Today, it’s a vast and complex field of research, theory and practice involving many aspects of human and AI communication.
Understanding the difference between human and machine decision-making will be the key here. How will the way that agents apply weight to factors like value, brand recognition or safety differ from the way we do?
But the most important piece of advice is to avoid the temptation to wait and see. The revolution is already underway, and as a growing number of AI service providers begin offering agentic platforms and apps, it could quickly become part of everyday life.
By the time this happens, the race could already have been won, either by an existing AI leader or some upstart that approached this challenge in a more agile, innovative way. Right now, though, there’s everything to play for and a hugely exciting opportunity for businesses wanting to be on the frontier of the next big wave.
