Jack Smyth, Jellyfish chief solutions officer for AI, writes that the next generation of the internet will be dominated by AI agents, leaving humans out in the cold and brands in a spin to appease these new tastemakers.
A week ago I saw a glimpse of the next web.
Much of it was familiar—browsers, webpages, address bars and plenty of stuff to click—with one thing missing.
People.
The next web may be built by and for a growing network of AI agents who browse and buy on our behalf.
Marketers need to reassess the very core of their strategy—the who, what, and where of marketing to find the new www for the next internet.
Who: Will Be Browsing and Buying?
Agents are everywhere in the trade press. If you’re still wondering what they are, you’re not alone. The best definition I’ve found is thanks to Amazon: “humans set goals, but an AI agent independently chooses the best actions it needs to perform to achieve those goals”.
Every single day your customers have a set of goals—from buying milk to browsing mortgages—that an agent could do.
AI-driven tools like OpenAI’s Operator and Instacart’s retail agents are stepping into the shopping journey at different steps based on capabilities and human curiosity.
Thankfully for marketers, agents aren’t replacing consumer choice. They’re streamlining it. Consumers may drive the process, but they increasingly rely on AI to filter options, validate choices, and recommend products based on predefined preferences.
So this doesn’t mean traditional marketing tactics become obsolete overnight, but it does require adaptation.
SEO, contextual targeting, and brand consistency become more important when a consumer’s AI assistant is choosing which results to show their owner.
The challenge is understanding how each agent curates choices and ensuring your brand remains visible within these evolving recommendation systems.
What: Do Those Models Think About Your Brand?
These agents will be powered by a handful of foundation models from familiar internet titans like Google. They’re the ‘brains’ behind the buying agent.
These models don’t have opinions in the traditional sense, but they do process and synthesize vast amounts of publicly available information.
Their “perception” of a brand is shaped by product reviews, media coverage, user discussions, and corporate messaging picked up through crawlers and content readining.
Unlike traditional sentiment analysis, which provides a snapshot of public opinion, AI models construct a more nuanced representation of brand positioning over time.
Understanding how these models interpret and compare brands is becoming as crucial as traditional brand tracking. This reality reinforces the importance of consistency in messaging and ensuring that positive narratives are well-represented in digital content ecosystems.
Where: Are You in Latent Space?
The brains behind the agents have a common structure. While marketers don’t have to keep up with every academic paper, there are critical concepts that help you understand how these models work.
One of those is latent space.
It’s an abstract concept in AI, representing how different ideas, brands, and themes are mapped in relation to each other.
Latent space is how AI models conceptualize brands based on the data they have processed. The implications are stark: brands have to differentiate themselves in the minds of consumers AND latent space.
Understanding your brand’s position in latent space can offer valuable insights into competitive positioning. If an AI model places your brand close to others in the same category, differentiation may be more difficult.
Conversely, if your brand is associated with distinct attributes—such as innovation, sustainability, or affordability—it is more likely to be surfaced in AI-driven recommendations aligned with those themes.
So in short: there’s good news and bad news.
The good news is many of the core tenets of marketing—distinctiveness, differentiation, consistency—will pay off in an agent economy.
The bad news is that it will be harder to gain an advantage in any of those areas as it becomes easier to track, compare, and copy successful tactics.
Consumers won’t be the only ones with agents, every agency and brand will have their own competitive intel network scouring for every incremental opportunity.
So enjoy the current version of the web while you can. The next iteration is going to be a wild ride.
