Brands rely on comparison content (listicles, X vs Y...) to promote their product. Intelligent AI models can see through thinly-veiled promotional content, which backlashes against the brand. In this article, we describe how to write comparison content that works in the AI age.

Your marketing pages are optimized for conversion. Your blog is optimized for engagement. Your API docs? They're optimized for clarity, completeness, and literal truth. That makes them the most powerful AI marketing asset you have—and most companies don't realize it.

When OpenAI launched ChatGPT's Shopping Assistant in December 2024, the tech press focused on "ChatGPT can buy things now." We saw something more important: a public demonstration of exactly how AI models make recommendations, and how brands should approach them more like a human influencer to persuade rather than a search engine to be gamed.

Enterprise software gets recommended by default—there's decades of web presence to crawl. But when a buyer asks ChatGPT for tools in your category and you're six months old, you're invisible. The model simply doesn't know you exist. This is the real problem most AEO tools miss: it's not about ranking, it's about discovery.

Most brands think AI visibility is a single problem: show up more often in AI responses. But recommendation actually requires clearing two sequential hurdles, each demanding its own strategy. The first is visibility—getting into the conversation at all. The second is specificity—being the recommendation when constraints tighten.

Last year, over 700 million conversations happened across AI chatbots about purchasing products. Those weren't searches—they were consultations. Something fundamental is shifting in how people make buying decisions, and it looks a lot less like googling and a lot more like asking a trusted friend for advice.

The shift from Google Search to ChatGPT as a personal, buying assistant means a shift from a ranked list of results to a single recommendation. For brands, this means that there is no "second place" when it comes to getting recommended by ChatGPT. Brands need to ensure that AI agents can find the specific information that they need to make a recommendation.

The web is quickly becoming infrastructure for bots and AI agents to interact with the world (gather information, take actions like purchases). At the same time, human attention is in the middle of the greatest migration since its first migration online. We can assume that these trends are going to continue.

When Google started losing share to ChatGPT, the first instinct was obvious: let's do SEO for ChatGPT. Track your "rank" on prompts, optimize for long-tail queries, measure your share of mentions. This is the horseless carriage stage—applying an old paradigm to a new technology. It's not wrong. It's incomplete.

For the last two decades, we've accepted an uncomfortable compromise: content that tries to please both humans and search engines ends up underwhelming both. Now there's a third constituency—AI models—and the compromise is untenable.

Marketing’s new audience is AI itself: people now start buying journeys by asking models like ChatGPT, Gemini, and Perplexity, which act as influential intermediaries deciding which brands to recommend. To win those recommendations, brands must treat models as rational, verification-oriented readers—using clear, specific, and consistent claims backed by evidence across sites, docs, and third-party sources. This unlocks a compounding advantage: AI systems can “show their work,” letting marketers diagnose how they’re being evaluated and then systematically adjust content so models—and therefore buyers—see them as the right fit.

Turns out that they are. Anthropic’s 2025 research shows advanced Claude models can sometimes detect and describe artificial “thoughts” injected into their own activations, providing the first causal evidence of genuine introspection rather than post-hoc storytelling—about a 20% success rate with zero false positives. The effect is strongest for abstract concepts and appears to rely on multiple specialized self-monitoring circuits that emerged through alignment training, not just scale. While this doesn’t prove consciousness, it demonstrates that leading models can access and report on parts of their internal state, with significant implications for interpretability, alignment, and how we evaluate future AI systems.