Semantic URL Slugs Boost AI Citation Rates by 21% Across All Six Major Engines

Searchable.com researchers Connor Lahey and Sean Fitzsimons ran a controlled study from October 17–26, 2026 across all six major AI engines. The finding: pages whose URL slugs semantically match the query are cited 21% more often than pages with identical titles but generic paths — IDs, dates, or category placeholders. The effect held on every engine tested.

URL structure has been a known SEO signal for over a decade. What's new is the magnitude of the effect on AI citation specifically, and the fact that it holds across engines with completely different citation mechanisms — including ChatGPT, which notoriously ignores organic search rankings for most signals.

What the Study Found

The study held page titles constant and varied only the URL slug. For example: a page titled “Best CRM for Small Business” at /products/item-4821 versus the same page at /crm/best-crm-small-business. Same content. Same title. Different slug.

The semantically matched slug won on every engine, every time. The effect sizes by engine:

• Perplexity: +31% — the largest gap. Perplexity is known to cite based on topical relevance signals; URL path appears to function as a strong semantic anchor for its retrieval system.
• Google AI Overviews and AI Mode: +22% — consistent with known Google behavior; URLs remain a ranking signal in organic search, and that signal appears to propagate to AI Overview sourcing.
• ChatGPT: +19% — notable because ChatGPT citation behavior is widely considered to be nearly independent of standard SEO signals. URL semantics appear to be an exception.
• Gemini and Microsoft Copilot: +15% — the lowest gap, still consistent and positive.

When both URL and title were semantically aligned with the query — not holding either constant — the citation lift jumped to +47% versus pages with neither aligned. Title alone (URL held generic) delivered +30%.

Within-site comparisons showed the same pattern: on sites with mixed URL structures (some descriptive slugs, some numeric IDs), the descriptive-slug pages were cited 27% more often for the same queries.

Why It Matters

1. The URL is a signal before the page is even loaded. AI retrieval systems process URLs as structured metadata. A slug like /blog/how-to-reduce-saas-churn signals topic, entity type, and query relevance before any content is parsed. A slug like /blog/post-3847 signals nothing.
2. This is the cheapest GEO fix that exists. Most GEO improvements require content rewrites, schema implementation, or link building. URL slug optimization is a one-time technical change per page — usually a CMS setting or a bulk redirect rule — with permanent citation benefit.
3. The engines that "don't use SEO signals" still respond to URL semantics. ChatGPT's +19% is the important finding here. The study adds URL slug to a very short list of signals that appear to influence ChatGPT citation across the board — alongside domain authority and entity disambiguation.
4. Generic slugs actively penalize you against competitors with descriptive ones. Below-average URL semantics means you're giving up 21%+ citation frequency every time you appear in the same candidate pool as a competitor who has clean slugs. That gap compounds across thousands of queries per month.

What to Do

1. Audit your current URL structure. Run a crawl or export from your CMS. Flag every page that uses:
   • Numeric IDs (/products/12847, /blog/post-2024-03-15)
   • Date-based paths without keywords (/2024/03/this-post/)
   • Generic category slugs (/resources/article, /news/update)
   • Hash fragments or query strings as primary identifiers (/page?id=482)
2. Prioritize high-value pages first. Start with: your top 20 most-linked pages, your product/service pages, and any pages currently ranking on page 1 for commercial queries. These are your most likely AI citation candidates — cleaning their slugs has the highest expected lift.
3. Set up 301 redirects before changing slugs.Any slug change without a redirect destroys the organic equity of the old URL. Implement the redirect first, confirm it's live, then update the slug. This is non-negotiable — the citation gain is not worth a ranking loss.
4. Write slugs for the query, not the internal content taxonomy. A page about enterprise SaaS pricing should be /enterprise-saas-pricing, not /pricing/tier-3. The slug should reflect how someone would search for the page, not how your team organizes files internally.
5. Align title and slug together.The +47% combined lift figure means title-slug alignment is where the real upside lives. Once you're rewriting a slug, review the title at the same time. Both should answer the query in the first 4–6 words.
6. Set a slug standard for new content now. The migration work on old content is a one-time project. The bigger win is preventing the problem from accumulating. Add a slug review step to your content publishing checklist. It takes 30 seconds per post.

Chad's Take

Every GEO conversation eventually arrives at the same uncomfortable question: “Why is a competitor with worse content ranking in AI answers?” URL semantics is one of the cleanest answers to that question — and the most embarrassing one, because it's fully within your control. You didn't lose the citation because the competitor wrote better content. You lost it because their slug said /best-crm-saas-2026 and yours said /products/p4821. Fix your slug structure this week. Not next quarter. This week.

Sources

• Searchable.com — URL Slug Semantics and AI Citation Rate study by Connor Lahey and Sean Fitzsimons; 6 AI engines tested (ChatGPT, Perplexity, Google AI Overviews, AI Mode, Gemini, Microsoft Copilot), October 17–26, 2026. searchable.com