Visible Date Signal: The Freshness Cue AI Checks Before Citing You

What this article answers

• Why do AI engines care about visible dates on my content?
• What format should publication dates be in for AI visibility?
• Should I show “last updated” dates or just publication dates?

Key takeaways

• Visible Date Signal checks for both human-readable dates on the page and machine-readable dates in Article schema.
• AI engines are recency-biased - content without visible dates is treated as potentially outdated, reducing citation probability.
• Show both “Published” and “Last Updated” dates. The update date tells AI (and users) that you actively maintain your content.
• Use ISO 8601 format (YYYY-MM-DD) in schema and a human-friendly format (“March 9, 2026”) in the visible layout.
• This is different from Content Freshness (which checks whether dates reference recent years) - Visible Date Signal just checks that dates exist and are visible.

The Date AI Looks For Before Citing You

Visible Date Signal carries 1.5% raw weight and checks whether your pages display clear publication and update dates both in human-readable text and machine-readable schema.

Every AI engine has a recency bias. When two sources provide the same information, the one with a visible, recent date wins. The one without a date gets treated as potentially stale - even if the content was published yesterday.

Visible Date Signal carries 1.5% raw weight inside the AEO Site Rank, in the Trust & Authority pillar. It measures a simple but critical thing: do your content pages display clear publication and update dates?

This is different from Content Freshness (ID 11, 4% weight), which checks whether your content references recent dates and events in the text body. Visible Date Signal only checks whether the date metadata exists - both visible to humans and machine-readable in schema.

The scorer looks for:

• Visible date elements on the page (text containing month/year patterns, HTML <time> elements)
• datePublished property in Article or BlogPosting schema
• dateModified property showing the content has been updated
• Consistency between visible dates and schema dates

A page with “Published March 2026” visible in the layout and matching datePublished: "2026-03-09" in Article schema scores well. A page with no date visible and no date in schema scores zero. A page with a date in schema but nothing visible on the page gets partial credit.

Why “Evergreen” Is Not an Excuse to Hide Dates

Removing dates to look evergreen backfires because AI engines guess conservatively about undated content, then prefer dated competitor pages even when yours is newer.

Some content teams deliberately remove dates from articles to make them look “evergreen.” The reasoning: if there’s no date, the content always looks current. This backfires completely in the AI era.

Without a date, AI engines must guess when your content was written. They often guess conservatively - treating undated content as potentially old. A competitor’s article from last month with a clear date will get cited over your undated article, even if your content is superior.

The “evergreen” approach worked when the only consumer was Google’s traditional search algorithm, which had other freshness signals to rely on. AI engines are less forgiving. They process content more literally and rely more heavily on explicit date signals.

The solution isn’t to fake dates - it’s to genuinely keep your content current and show when you did. A “Published January 2025, Updated March 2026” date tells AI three things:

1. This content has been around long enough to be trusted
2. Someone actively maintains it
3. The information reflects 2026 reality

That combination of maturity and freshness is the strongest possible date signal.

Implementing Visible Date Signals

Use the HTML time element with ISO 8601 datetime, add datePublished and dateModified to Article schema, then keep visible dates and schema dates synchronized across updates.

1. Add visible dates to your article template

<article>
  <header>
    <h1>Article Title</h1>
    <div class="meta">
      <time datetime="2026-03-09">Published March 9, 2026</time>
      <time datetime="2026-03-09">Last updated March 9, 2026</time>
    </div>
  </header>
</article>

Use the HTML <time> element with a datetime attribute in ISO 8601 format. The visible text can be in any human-friendly format.

2. Add dates to Article schema

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Article Title",
  "datePublished": "2026-03-09",
  "dateModified": "2026-03-09"
}

Both datePublished and dateModified should be present. If the article hasn’t been modified since publication, use the same date for both.

3. Show “Last Updated” on revised content When you update an article, update the visible date and the dateModified schema property. This is the strongest freshness signal you can send - it proves active maintenance.

4. Be consistent If your visible date says “March 9, 2026” but your schema says “2025-12-01,” AI gets conflicting signals. Both sources must agree.

5. Don’t backdate or forward-date Artificially changing dates to appear more recent is a short-term tactic that AI engines are increasingly sophisticated at detecting. Use real dates. If your content is from 2024, show 2024 and add an “Updated 2026” notice when you revise it.

Start here: check your blog template. If there’s no visible date, add datePublished and dateModified to both the page layout and the Article schema. This is a template-level change that fixes every article at once.

Visible date signals must appear in three places to lift the AEO Rank freshness check.

Score Impact in Practice

Sites with both visible dates and matching Article schema score 8-10/10, while schema-only or visible-only sites earn partial 4-5/10 credit and undated content scores zero.

Visible Date Signal carries 1.5% raw weight in the Trust & Authority pillar. Sites with both visible dates and matching datePublished/dateModified in Article schema score 8-10/10. Sites with dates in schema but nothing visible on the page score 4-5/10 (partial credit). Sites with no dates anywhere score 0.

The 2% weight understates the practical impact because dates influence how AI engines evaluate other criteria. Content Freshness (4% weight) checks whether your content references recent dates in the body text, but it starts by looking for a publication date to anchor the freshness assessment. Without a Visible Date Signal, the Content Freshness scorer has no baseline - and a page with no baseline date tends to be treated as potentially stale.

In our audits, approximately 35% of sites have no visible dates on their content pages. This is particularly common among SaaS companies that intentionally strip dates to make pages look “evergreen.” Among those sites, the average Visible Date Signal score is 1.2/10. Sites that show both publication and last-updated dates average 8.5/10. The fix is a template-level change - add the date display to your article template once and every existing article gets the signal immediately.

Where Sites Lose Points

Dates in schema but not on the page, dates on the page but not in schema, and mismatched dates between the two each cost partial credit on this criterion.

Dates in schema but not on the page. This is the most common partial failure. The developer adds datePublished to Article JSON-LD because a checklist told them to, but the design team didn’t include a visible date in the page layout. AI systems that parse schema get the signal, but systems that extract from visible content don’t. You get partial credit at best.

Dates visible on the page but not in schema. The reverse problem - a “Published January 15, 2026” line in the article header, but no datePublished in Article JSON-LD. This is worse than the first case because schema is the higher-confidence signal. Visible text dates require AI to parse natural language date formats, which introduces ambiguity.

Inconsistent dates between schema and visible content. The visible page says “Published March 2026” but the schema says “datePublished: 2025-12-01.” This conflicting signal is worse than having no date at all because AI engines treat the inconsistency as a trust failure. Always ensure schema dates match visible dates exactly.

Using only datePublished without dateModified. A page published in 2024 with no dateModified tells AI the content hasn’t been touched in two years. If you’ve actually updated the content, adding dateModified with the current date sends the freshness signal. If you haven’t updated it, the missing dateModified is accurate - but it means your content looks stale compared to a competitor who recently updated theirs.

Relative dates instead of absolute dates. “Posted 3 months ago” becomes meaningless the moment AI caches the page. By the time a user sees the AI’s cached version, “3 months ago” could be a year old. Always use absolute dates that don’t decay: “Published January 15, 2026.”

How AI Engines Evaluate This

Every major AI engine has a recency preference, and visible date signals are the primary input for that preference. The specifics vary by engine.

ChatGPT checks for datePublished and dateModified in Article schema first. When both are present and recent, ChatGPT treats the content as current. When only datePublished exists, ChatGPT uses the publication date as the freshness anchor. When neither exists, ChatGPT falls back to inferring dates from the content body - looking for year references, event mentions, or other temporal markers. This fallback is unreliable and often produces conservative (older) date estimates.

Claude places high weight on dateModified specifically. A page with dateModified: "2026-03-01" tells Claude that someone actively reviewed and updated this content recently. Claude treats the presence of dateModified as a maintenance signal - evidence that the content is actively curated rather than published and abandoned. Pages with only datePublished get lower freshness confidence, and pages with neither get treated as undated content with unknown freshness.

Perplexity is the most recency-biased engine. When building real-time answers, Perplexity explicitly prefers recent sources. A page from 2026 with a visible date gets prioritized over an undated page - even if the undated page has superior content. Perplexity also respects the HTML <time> element directly, using the datetime attribute for unambiguous date parsing. Using <time datetime="2026-03-09"> instead of plain text gives Perplexity a machine-parseable date that doesn’t require natural language processing.

Google AI Overviews uses datePublished and dateModified from Article schema as inputs to its freshness ranking system. Pages with recent dateModified values are eligible for time-sensitive queries (“best live chat tools 2026”) while pages with old or missing dates get filtered out. The visible date on the page also feeds into Google’s “date snippet” in search results, which influences click-through rates that indirectly affect AI Overviews source selection.

External Resources

• Schema.org Article datePublished - https://schema.org/datePublished
• Schema.org dateModified - https://schema.org/dateModified
• HTML time Element - https://developer.mozilla.org/en-US/docs/Web/HTML/Element/time
• Google Search Central: Article Structured Data - https://developers.google.com/search/docs/appearance/structured-data/article

Related topics

• content-freshness
• structured-data
• content-velocity