Domains that update their sitemap lastmod timestamps on a regular cadence are cited 31% more frequently by retrieval-augmented engines than domains with static or absent lastmod values. The freshness signal is measurable, controllable, and largely ignored.
Freshness is a first-class citation signal in retrieval-augmented AI engines. Perplexity, Bing Copilot, and SearchGPT all apply recency weighting when selecting which pages to cite. A page updated three weeks ago consistently outperforms an equivalent page last touched in 2022 — even when the 2022 page has higher inbound link authority.
This is not true for base-model citations (GPT or Claude without web browsing). But retrieval-augmented engines now represent a growing share of AI answer traffic, and the freshness gap between maintained and unmaintained sites is widening.
The lastmod signal
The sitemap.xml lastmod field tells crawlers when each URL was last meaningfully modified. Retrieval-augmented engines use this as the primary freshness input — not the page's HTTP Last-Modified header, not the on-page published date, not inferred signals from on-page content. Sitemap lastmod is the authoritative freshness claim.
The implication: you can improve your freshness score without publishing new content. Update your sitemap lastmod dates when you make genuine improvements to existing pages — correcting facts, adding a new section, updating statistics. This signals to retrieval systems that the page reflects current information.
What you should not do: update lastmod without making substantive changes to the page. Search engines validate lastmod against actual content changes. Pages with fraudulent lastmod values are penalised — the models learn to distrust the signal from that domain.
The freshness gap by content type
Freshness weighting is not uniform across query types. These are the query categories where recency has the highest citation impact:
Pricing queries: "how much does [product] cost" — models heavily prefer recently-indexed pages because prices change. A pricing page with a 2024 sitemap entry is cited over a technically superior 2022 page.
Benchmark and comparison queries: "best [category] tools in 2026" — recency is implied by the year in the query. Pages updated in the current year have a structural advantage.
How-to and tutorial queries: "how to [task]" — moderately freshness-weighted. A 2021 tutorial on a stable technology ranks well; a 2021 tutorial on a fast-moving API stack does not.
Entity and background queries: "what is [company]" — low freshness weighting. Wikidata and Wikipedia are treated as authoritative regardless of their last-update timestamp.
Citation data: the lastmod effect
Across the domains tracked in this week's index, domains with complete sitemap lastmod values updated within the last 90 days received citation rates 31% higher in retrieval-augmented engines than comparable domains with missing or static lastmod. The gap is largest for pricing and how-to content — the query types most sensitive to recency.
The fix is mechanical. Automate your sitemap lastmod values to reflect actual content change timestamps. Most CMS platforms do this natively but have it disabled by default. For custom Next.js sites, generate sitemaps dynamically with the current timestamp applied only to recently-modified pages.
Freshness and the Article schema dateModified field
Article schema includes a dateModified field that retrieval engines use as a secondary freshness signal, independent of sitemap lastmod. For blog posts, case studies, and long-form pages: update dateModified in your JSON-LD whenever you make substantive edits. The combination of accurate sitemap lastmod and correct dateModified is the strongest freshness signal available in on-page schema.
See the structured data playbook for Article schema implementation. Keep dateModified current; an outdated dateModified paired with a current sitemap lastmod creates a mixed signal that some engines penalise.
What this means for your maintenance cadence
The minimum viable freshness practice: review your 20 highest-traffic pages quarterly. Update any outdated statistics, pricing, or feature descriptions. Push updated sitemap with current lastmod values. Update Article schema dateModified where applicable.
This takes 2-4 hours per quarter and produces measurable citation rate improvements within 2-3 weeks in retrieval-augmented engines. Scan your domain to see your sitemap freshness score and which pages are dragging your citation rate down with stale lastmod values.
Put this into practice
See how your domain scores on the signals covered in this edition. Veezow runs a free AI visibility scan — robots, sitemap, structured data, bot access, and off-site presence.
Run a free scan →