Clean, descriptive URLs may support Ai retrieval, but existing sites should avoid risky restructures and apply the principles to new content.
URL structure is hardly the most glamorous part of SEO. Along with XML sitemaps, redirects and canonical tags, it’s one of the reasons why SEO might be better off with the IT department.
How To Design URL Structures For AI Retrieval, Not Just Rankings on Search Engine Journal article reminds us of a useful point: URLs are not just addresses. A clear URL can tell search engines, Ai systems and human users what a page is likely to be about before anyone has even read the page.
We should not redesign every website around Ai retrieval, however. URL changes can break things. But it is a good reason to think more carefully about new content.
So what is the old advice? URLs should be descriptive and reasonably short. They should avoid internal jargon and random numbers. That clarity helps people. In Ai search, where systems may be pulling passages from different sources and deciding what to cite, those small signals could help reinforce what a page is really about. A good URL will not turn an average article into an authority. But it can support the case the page is already making.
When publishing new articles, we should ask whether the URL tells a clear story. Could someone understand the likely content from the URL alone? That’s the goal.
For existing content, the bar should be higher. We should only change URLs where there is a clear commercial or structural reason. Keep the current site stable unless there is a strong reason to change it. Audit important pages where risk is low and reward is clear. For everything new, apply better principles from now on.