To say the least, 2025 has been a transformative year for SEO. Here are the five most important things I’ve learned or refined my thinking on this year that are shaping the advice I’m giving clients heading into 2026:
1. AI is a closed ecosystem
AI search is largely a closed ecosystem where traffic is a secondary outcome, not the primary goal. Like it or not, optimization now is about being cited, summarized, and mentioned within the AI response itself. AI visibility tools are the best option to monitor this at present and we need to make sure the prompts we’re tracking are representative of what prompts people are actually entering.
2. Bots & agents are now a primary audience
Humans are visiting websites less directly and increasingly relying on bots or agents to visit on their behalf. In other cases, humans are visiting along with bots or agents. User journeys are really different now. That means bots and agents need to be treated as a primary audience. Metrics, goals, and optimization strategies all need to reflect this new reality.
3. Audit for bot & agent access
Do your technical foundations support agents and AI bots? Are you clear on what content bots should and should not use for training?
There is a lot to do here, including:
- Check your server logs or use tools like Dark Visitors to understand how your site is being crawled and by which bots.
- Make sure your hosting provider’s firewall isn’t “helpfully” blocking AI bots.
- Confirm your robots.txt rules are intentional.
- Start exploring protocols like Model Context Protocol (MCP) or Agent-to-Agent (A2A).
4. Server-side rendering matters more than ever
Most AI bots, including ChatGPT-User, do not currently execute JavaScript. If you want AI bots to reliably find your content, it must exist in the initial server-side HTML. As part of this, don’t forget to include internal links in server-side HTML as well so that bots can find links to other pages. Tools like the Chrome extension View Rendered Source make this easier to check and monitor.
Closely related: optimizing Time to First Byte (TTFB) makes it easier for bots to fetch that server-side code in the first place.
5. Shift to a vector mindset
AI relies on vector search to match intent and semantic relationships. Technically, that means stronger internal linking, clearer taxonomies, and more focus on site structure to reinforce your brand’s relationship with specific entities (real-world concepts). Semantic HTML (like <article> and proper <h> tag usage) and schema may also help bots identify and extract content (more experiments and testing are needed).