Sitemap XML: how do you update it for GEO?

Sitemap : XML sitemaps still matter, but LLMs treat them differently. Auto-generate on every deploy, segment by language and content type, submit to Bing Webmaster + IndexNow. Freshness (<lastmod>) outweighs size.

Sitemap XML: how do you update it for GEO, in real use?

The classic sitemap listed URLs for Googlebot. In 2026 the audience includes OAI-SearchBot, PerplexityBot, Bingbot and Applebot. My PROEMA principle: segment the sitemap. A sitemap-index pointing to sitemap-fr.xml, sitemap-en.xml, sitemap-nl.xml, sitemap-blog.xml, sitemap-faq.xml. That granularity helps bots prioritise. On expertcafe.be (5 languages, ~10,000 URLs) segmentation cut the observed full re-crawl time for Bingbot.

Tags that still matter and tags to drop

"The lastmod tag must be real, not always set to today, otherwise bots learn to ignore it and treat it as noise," stresses Lorenzo Eeman, founder of PROEMA. Frequency (<changefreq>) is obsolete, ignored by Google since 2023. Priority (<priority>) too. What counts: completeness (every canonical URL), lastmod freshness, strict consistency with hreflang tags. Common PROEMA finding: a sitemap that lists pages with noindex or 301 redirects, AI bots penalise that incoherence.

IndexNow and the Cloudflare build hook

Pushing via IndexNow (a protocol adopted by Bing, Yandex, Naver and Cloudflare) accelerates indexation based on Microsoft 2024 benchmarks, the API is free, takes 5 minutes to wire, and publishes content changes immediately to partner engines. For Cloudflare Pages I use a build hook that regenerates the sitemap on every deploy and pings IndexNow right after. Quarterly PROEMA audit: orphan URLs, 404s, redirect chains, language duplicates, and a check that robots.txt carries the Sitemap: directive as an absolute HTTPS URL. Sources: developers.google.com/search/docs/crawling-indexing/sitemaps, indexnow.org, Bing Webmaster blog.