We came away from this year’s MozCon with pages of notes on AI search, attribution, crawling and agentic commerce. But the more we compared notes afterwards, the more it felt like every talk was pointing to the same underlying shift.
Marketing is no longer just about influencing people. Increasingly, it’s about influencing the AI systems making decisions on their behalf.
One phrase captured that shift particularly well: Machine Media. The idea that AI increasingly gathers, interprets and curates information before a human ever reaches a website.
And it’s already showing up in the data. Cloudflare found that while human web traffic grew by just 3.1% during 2025, AI bot traffic increased by 187%. Even more striking was the growth of agentic traffic, where autonomous systems retrieve information and act on a user’s behalf. That grew by 7,851% in a single year.
Why Traditional Marketing Models Are Breaking
This changes far more than search.
As AI agents become capable of completing tasks on behalf of users, discovery and checkout collapse into a single interaction. Rather than recommending products and sending users to a website, agents can compare options, make decisions and complete transactions directly within their own interfaces.
That creates a problem for marketing teams. Attribution models were built around observable customer journeys, but those journeys are becoming increasingly invisible. If an AI agent researches, evaluates and purchases without sending a visitor to your site, many of the signals marketers have relied on for decades begin to disappear.
The underlying infrastructure is evolving just as quickly. Emerging standards such as OpenAI’s Agentic Commerce Protocol and Stripe’s Shared Payment Token allow AI systems to query inventory, compare products and execute payments directly. In this environment, content is no longer simply designed to attract visitors. It becomes structured information that machines consume in order to make decisions.
The question therefore changes. It is no longer How do we get someone to click? It becomes How do we become the brand an AI chooses?
The Answer Isn’t More Content. It’s More Context.
The instinctive response to declining traffic is often to produce more content.
But AI systems don’t choose brands based on volume alone. They choose brands that are most relevant to the individual user they’re serving.
Google has already demonstrated this direction through its Personal Intelligence systems, which combine explicit context, such as a user’s stated preferences, with implicit signals drawn from search history, location, Gmail and previous behaviour.
An iPullRank study illustrates just how powerful that context has become. Researchers compared responses between a blank Google account and another seeded with brand signals across Gmail and Google Photos. In the control account, brand visibility remained virtually unchanged. In the personalised account, brand appearances increased from 21.9% to 62.3%.
The implication is significant. High-quality content remains essential, but content alone is no longer sufficient. If AI systems cannot connect your brand to a user’s context, they are increasingly unlikely to recommend you in the first place.
The competitive advantage is shifting from publishing more content to becoming more relevant.
Introducing Relevance Engineering
This is why marketing needs a new operating model.
We call it Relevance Engineering: the systematic process of structuring a brand’s content, data and digital presence so AI systems can confidently understand, retrieve, cite and ultimately choose it.
Unlike traditional SEO, Relevance Engineering doesn’t focus on optimizing for a single search engine. It recognizes that every AI platform retrieves and evaluates information differently. Claude, for example, relies heavily on Brave’s search index, while other models draw from entirely different retrieval systems. Success depends on ensuring your brand is consistently legible across all of them.
That also changes how performance should be measured. Rankings alone are becoming less meaningful as AI Overviews and conversational interfaces replace traditional search results. Instead, marketers need to understand whether machines can access their content, whether models cite their brand and whether those citations ultimately influence commercial outcomes.
A practical measurement framework should therefore include three layers:
- Input metrics: Can AI systems crawl, understand and retrieve your content through structured information, schema and clearly expressed answers?
- Visibility metrics: How often is your brand cited, recommended and accurately represented across AI platforms?
- Business metrics: Does machine visibility translate into leads, purchases and revenue, whether or not a website visit occurs?
What Marketing Leaders Should Do Now
Relevance Engineering isn’t another marketing channel. It’s a coordination challenge.
No single team owns the capabilities required to succeed. Structured data sits with engineering, digital PR builds authority, CRM shapes customer context, and content teams create the information AI systems retrieve. Without coordination across these functions, brands risk remaining invisible to the very systems increasingly making purchasing decisions.
That makes three priorities particularly important over the coming quarter:
- Establish clear ownership for AI visibility across marketing, engineering and product teams.
- Expand reporting beyond rankings to include AI citations, share of model and agentic conversions.
- Invest in structural improvements such as schema, open protocols and machine-readable content before simply increasing content production.
The Bottom Line
For years, digital marketing has been built around influencing human attention.
The next phase will be built around influencing machine decisions.
That doesn’t make content less important. It changes what content needs to achieve. The brands that succeed won’t necessarily publish the most. They’ll be the ones whose products, expertise and authority are easiest for AI systems to understand, trust and recommend.
That’s the role of Relevance Engineering, and why it is rapidly becoming a foundational marketing discipline for the AI era.
