How to optimize for AI agents
If you’re freaking out about optimizing for AI agents, you are already behind.
This week’s newsletter is sponsored by North Star Inbound and Botify.
If you’re freaking out about optimizing for AI agents, you are already behind.
The social media hype pool is deafening right now. ChatGPT is killing search traffic. Atlas killed Chrome. AI Mode is cannibalizing Google. Perplexity’s the new Google.
[Sponsored by Botify]
Stop guessing whether AI visibility has an impact!
You can’t “rank” in AI search — every answer is unique. Traditional tracking doesn’t work, and guessing isn’t a strategy.
Built on first-party data and backed by powerful AI-powered insights and automation, AI Visibility measures your actual brand impact in AI answers. You get measurement, intelligence, and action — all in one platform.
How are you showing up in ChatGPT, Perplexity, Google AI Mode, and more? Get early access to AI Visibility in beta, new from Botify.
Every marketing blog is pushing an “AI SEO checklist” and playbook to “rank on AEO”. In all of my client calls, everyone is obsessed with one question: how do I optimize for AI and AI agents? Sometimes they use words they have heard on LinkedIn, like GEO or AEO, but I am frequently surprised to hear new ones. (Comment with your favorite ones.)
I firmly believe that not much needs to change if you’ve been building for users all along. An AI agent is essentially a user with an autonomous interface, seeking the same goal as a human researcher: precise, accurate, and well-organized answers that effectively solve what they are searching for. If your website and content deliver value to people, they also provide for AI.
The sites panicking now are the ones that spent years gaming the system, stuffing keywords and chasing backlinks while optimizing for crawlers instead of comprehension. These sites are now facing LLMs that train themselves to ignore spam efforts that are not user-centric.
What AI agents need
When an AI agent finds your content, it’s doing homework on behalf of a user who has a question, a problem, or a decision to make, and the agent’s job is to come up with the answer. It’s looking for the same things a human would: answers.
If you’ve been doing SEO with users in mind, the agent is just a more sophisticated user.
Just like you would use an effective word (not just a keyword) strategy for humans, agents need the same. Agents might act autonomously, but they are still just running Google (or other) searches. With agents, it's even more important to answer their questions thoroughly, not just to rank for them, because the agent will develop better top-line search queries than a human.
Agent Best Practices
Don’t focus on word count and keyword density. Write for clarity instead because AI can spot the difference between content that genuinely explains something and content that’s just filling space. If you’re padding to hit 2,000 words because your SEO agency told you to, just don’t.
Organize your content so it's easy for humans to read, with information structured in a way that makes sense to HUMANS. Ensure each section flows naturally into the next and is well supported. (AI writing tools can help, but make sure you proofread.)
Back up your claims with real sources; citations matter tremendously in an AI world. Unlock traditional algorithms which use linear algorithms (website→ website), AI looks at context and content more holistically. (website ←→ website)
Content needs to be scannable without sacrificing depth, allowing readers to grasp key points by skimming and dive deeper into any section that matters to them. If a bulleted list addresses the question, that’s all you need. It doesn’t need to be text walls that require scrolling through a page to find the answer. Too many sites bury answers in a wall of text because they want more keywords in the content. That stopped working years ago, but now it actually has adverse effects, frustrating humans and AI alike.
Use structured data for information with a clear structure, such as events, products, recipes, or business information, as schema markup makes this information machine-readable. It may or may not help, but it certainly doesn’t hurt.
Keep your content fresh by regularly updating it, rather than letting it languish in your CMS for years. If your industry is rapidly evolving or new research contradicts your previous content, consider updating it or removing it, as outdated information can be more harmful than no information at all.
If you want the content to be considered authoritative, ensure you demonstrate genuine expertise rather than superficial research that any intern could accomplish in an hour with Google. If you're only regurgitating AI content, the AI will recognize itself and not value it. According to this Graphite study, more than half of internet content is generated by AI, yet LLMs ignore much of it.
Agents are not dominant and may never be
Marketers have always declared a new change that will disrupt everything, but it is essential to be realistic. Mobile-first indexing was going to ruin everything, voice search was the end of written content, Core Web Vitals were make-or-break for every site, and now it’s AI agents coming to destroy your traffic.
Some industries are feeling the impact of the AI shift more intensely than others, while others won’t be affected. If your industry is unlikely to have agents, keep the focus on the humans.
Finance, legal, and even health verticals might be one of those places where agents will gather information to compile reports and background research, so if this is your vertical, think about all the users and agents you need to address. Anything highly technical is a great place to invest resources and time, ensuring that agents can find it and use it as source material for their outputs.
Unlike with traditional SEO, where you might ignore technical documentation and developer resources as pages that should be indexed, these are already heavily used by AI coding assistants, so unclear or outdated docs get skipped in favor of sources that actually explain what developers need to know.
Focus on the user who will access your content
Your AI strategy should be identical to your plan for humans; if not, you never really had a user strategy in the first place. You had an algorithmic approach, and those always came with an expiration date baked in as algorithms got smarter.
Change in search is always a constant. Some of these changes actually matter, and some don’t, but the pattern’s always the same: if you built for users from the start, you adapt without breaking a sweat, and if you built for algorithms, you’re always scrambling to dig out of a traffic cliff. (Many people are trying this with LLMs by focusing on Reddit as a workaround. )
The technology and interfaces change, but people’s fundamental needs remain relatively unchanged. People want answers to what they search for. Whether that information reaches them through Google, an AI agent, or whatever gets invented next, the content that wins is always the one that actually delivers on that need, not just pretends to.
Stop worrying about the next big thing and focus on serving the actual needs users have been expressing all along. Follow the traditional SEO playbook that has worked for years, and cater to the most advanced Google searcher you have ever met: That is an agent.
Brought to you by North Star Inbound—the sales enablement SEO agency.
Drive High Intent Leads with SEO.
What happens when you combine best-in-class SEO with conversion optimization?
BigRentz’s traffic increased by 186% (85k), added 1950 conversions in 12 months.
Self Financial’s traffic increased by 50k/month, and added 685 new customers.
Lastly, Secure Data added 1968 phone calls.
North Star Inbound’s SEO strategies earn leads, conversions, and revenue..
Book a call for a free content audit and 10% off any engagement.
Intrigued? [Learn more here]
Thanks for writing this, it clarifies a lot. What if building for users means we still need to adapt to how AI interpets user needs?