**TL;DR** — Across 30 client sites through June 2026 we audited a structural choice that lives in whether the answer sentence pins itself to a time: whether the passage that answers a query carries an explicit **temporal scope** — an "as of" clause, a year, a stated version ("As of 2026, Google renders the answer card above the first organic result.") — or is written as an **undated timeless claim** that states the fact as if it holds always ("Google renders the answer card above the first organic result."), and whether stamping the sentence with its time changes how often the AI Overview lifts it into the card. Across 7,610 cited-passage events we joined each cited sentence to whether it carried a temporal scope. The headline is that temporal scope is a real citation lever, and it is really a freshness-verifiability lever wearing a date. An answer sentence carrying an explicit "as of" date was cited 1.7× more often than a matched undated sentence making the same claim on the same query, and the gap widened to 2.3× on queries about facts that visibly change year to year. The strongest predictor was recency-checkability — a sentence whose date a reader could see was current was lifted far more than one whose currency the reader had to assume. The second was scope-in-the-sentence — a date that sat in the lifted sentence beat one parked in a "last updated" line the composer did not lift. The third, and the warning, was stale-stamping — a visible date more than a year old was cited less than the same claim left undated, because the stamp advertised the staleness the undated version hid. One change — adding an explicit "as of" scope to answer sentences on time-sensitive queries — lifted cited-passage rate by 17% on the affected sites over a 30-day follow-up.
Why we ran this audit
The AI Overview composer lifts a single sentence and drops it into a card as the answer to a query, and it does so at a moment in time — the user asks today, and the card answers as if today. A whole class of answer sentences states a fact with no time attached: "the free tier includes 10,000 requests", "the algorithm favours dwell time", "the tax rate is 21%". Read inside an article dated last week these feel current, because the reader infers the date from the page; lifted cold into a card, an undated claim is a fact with no clock on it — the composer cannot tell whether it was true in 2019 or true this morning, and a fact that might be stale is a worse answer than one stamped current. We had spent weeks on the shape of the answer sentence — its polarity, its quantity, whether it defined its term — and whether the sentence carries its own time is the natural next structural variable, because time is the one dimension a fact silently loses when it is lifted out of a dated page.
The second motivation was a drafting habit that trusts the page date to carry the currency. A page publishes with a visible date, states its facts in the timeless present, and never restamps them in the prose — because the writer knows the reader can see the page was updated last month, and adding "as of June 2026" to every sentence reads as clutter to someone reading top-to-bottom. But the composer does not carry the page date into the card; it lifts the prose sentence, and the prose sentence surrendered its currency to the masthead date beside it. We needed to know whether folding an explicit temporal scope into the answer sentence — at the cost of a date that reads slightly redundant next to a fresh page date — bought the citation, because if it did, the fix is nearly free: move the currency from the masthead into the sentence the composer would lift.
How we ran the measurement
30 client sites — 11 SaaS, 6 publisher, 8 B2B services, 5 DTC — each with a fixed 200-query basket of its real in-market queries, deliberately weighted toward queries whose best answer turns on a fact that changes over time (pricing, limits, rankings factors, rates, version-specific behaviour). Twice daily through June 2026 we captured every AI Overview card, and for cards citing a client page we identified the specific lifted sentence and classified its temporal scope: dated (an explicit "as of", year, or version sits in the sentence), timeless (the claim stated in the plain present with no time marker), or off-sentence (the currency lives in a "last updated" line or masthead date but not in the prose). For each cited sentence we built a matched control: a comparable sentence on a similar query whose temporal scope differed but whose underlying claim was the same, so the comparison was dated-vs-timeless rather than good-page-vs-bad-page. The cited cohort was 7,610 events.
Two normalisation moves matter. We scored shape on the sentence as it would be lifted — alone, with no surrounding context — because that is the unit the composer extracts, and a timeless claim that reads as obviously current inside a page dated this week reads as a clockless assertion in the card. And we matched on sentence citability before comparing shape — we paired each cited sentence with a control our existing cited-paragraph rubric scored as equally liftable (concrete, on the query, factually complete), so the effect we attribute to the temporal scope is not just the dated pages being better maintained overall. The 1.7× and 2.3× figures are from those matched comparisons, not raw averages.
The shape of the temporal-scope pattern
The flat headline first. A sentence carrying an explicit date was cited more. An answer sentence stamped with an "as of" scope was lifted 1.7× more often than a matched timeless sentence making the same claim on the same query, and 2.3× more on queries about facts that visibly change year to year. The effect held through the quality match and the citability control: among sentences our rubric scored as equally liftable, the dated ones were lifted far more than the timeless ones. The composer behaves as though it prefers a claim it can date — one that tells it when the fact was true — over one that asserts a fact with no clock, because a card that says "As of 2026, the free tier includes 10,000 requests" answers the question and vouches for its own currency, while a card that says "the free tier includes 10,000 requests" answers with a fact that might be two years stale.
The most decision-relevant cut was that this is about the currency being checkable, not about a date being present. We tested whether the win came from any date appearing or from the date actually vouching for the claim being current, and the second was the whole story: a sentence with a decorative date that scoped nothing ("Founded in 2019, we find the free tier includes 10,000 requests") was cited no better than a bare timeless one, while a sentence whose date scoped the claim itself ("As of 2026, the free tier includes 10,000 requests") was lifted far more. The temporal scope wins when it makes the claim datable. Stamp the fact, not the sentence around it.
Driver one: let the reader see the fact is current
The single strongest predictor was whether a reader could see the claim was current. Holding the claim constant, a dated sentence was lifted at 1.7× the rate of a timeless one, rising to 2.3× where the fact plainly moves year to year. The composer extracts a sentence and puts it in front of a user who did not read the page; a sentence that says "As of 2026, Google renders the answer card above the first organic result" answers the question and hands the reader a currency they can trust, while one that says "Google renders the answer card above the first organic result" answers with a fact that might describe a layout retired two years ago — and a card that asserts an undatable claim is, to the composer, a worse answer than one that dates itself. A human reading the page in order sees the byline date; the reader in front of the card sees only the sentence, and the composer rewards the sentence that carries its own clock.
We ran a structural test on 28 answer sentences across 15 clients, each a strong claim on a time-sensitive query where the sentence stated the fact timelessly on a page that was in fact freshly maintained. We rewrote each to fold an explicit "as of" scope into the sentence, changing no underlying claim — only lifting the currency that lived in the masthead into the prose the composer would extract. Over the 45 days that followed, 20 of the 28 sentences began being lifted on at least one query where the timeless version had been skipped. The lever was not new information; it was moving the currency the page already had into the single sentence the composer would lift, so the sentence vouched for its own freshness.
Driver two: the date has to be in the sentence, not the footer
Holding dated constant, the second driver was where the date sat. A scope that lived in the answer sentence beat one parked in a "last updated" line, a footer, or a schema date the composer did not lift. The reading consistent with the data is that the composer lifts one self-contained sentence and does not reach up to the masthead beside it, so a page that put its currency in a "last updated" stamp and its prose in the timeless present gave the composer a sentence that still read as clockless — the date was on the page but not in the unit that got extracted. A page-level date is invisible to the lift; only the scope inside the sentence rides along into the card.
We ran a structural test on 18 off-sentence pages across 11 clients, each of which carried a fresh "last updated" date but stated its answer in the plain timeless present. We rewrote each to state the scope in the sentence itself, keeping the page date as the fuller signal, changing no facts. Over the 60 days after the change, 13 of the 18 sentences improved their cited-passage rate. The two drivers compound: a dated claim beats a timeless one, but only if the date is in the sentence — the sentences that won stated their scope in the prose the composer lifts, not in the footer it leaves behind.
Driver three: stale-stamping, and the date that hurt
The third driver was the warning. A temporal scope helps only when it dates the claim as current, and stamping a visible old date onto a claim backfires against leaving it undated. A sentence like "As of 2023, the free tier includes 10,000 requests" — a three-year-old scope on a fact the query wants answered today — was cited less often than the same claim left timeless, and on 7% of audited pages the composer paired the stale-stamped sentence with a competitor's current-dated one, so the citation was lost rather than shared. The reading consistent with the data is that an explicit old date advertises the staleness the timeless version hid; a visible 2023 tells the composer the fact may have moved since, where the undated claim at least left the currency ambiguous. Stamp the date only when it is current, and restamp it when you refresh the fact — a scope is a freshness signal only while it is fresh.
We confirmed this on 15 sentences across 9 clients where an earlier optimisation pass had hard-coded a now-old year into the answer. We rewrote each to carry the current scope where the fact had been re-verified, and stripped the scope back to a maintained timeless present where we could not re-verify it, matching the stamp to what we could actually vouch for. Over the following 45 days the current-dated sentences regained their citation while the honestly-timeless ones stopped advertising an old year. The actionable rule is blunt: a current date beats a timeless claim, but a stale date beats nothing — stamp the fact only as fresh as you can stand behind, and refresh the stamp when you refresh the fact.
What changed in our content checklist
Three changes. We added a temporal-scope pass for answer sentences: before publishing, we read each section's lead answer sentence alone and check that any time-sensitive fact it asserts carries an explicit "as of" scope, not a bare timeless present — because the composer lifts the sentence whole and hands it to a user who cannot see the byline date, so an undated fact is a claim with no clock on it. We added a location check to the same pass: the date lives in the sentence, not only in a "last updated" line, so when the composer lifts the sentence the currency rides along. And we added a freshness guard: the stamp is only as old as the fact is re-verified, so a current scope vouches for the claim and a stale one never advertises a year we can no longer stand behind.
We dropped one habit. For years our style had been to state facts in the timeless present and let the page date carry the currency — on the belief that stamping every sentence read as clutter to a reader who could see when the page was updated. The audit removes that default for the answer sentence on time-sensitive queries: the one sentence the composer would lift travels without its byline, and a timeless claim lifted alone asserts a fact with no clock. So the bare timeless present left our playbook for time-sensitive answer sentences — we now fold an explicit scope into the lead answer sentence, accepting that it reads slightly redundant next to a fresh page date, because it is built to carry its currency to the reader who cannot see the page date at all.
- 01Scope the fact to a time. A sentence carrying an "as of" date was cited 1.7× more than a timeless one making the same claim — 2.3× on facts that visibly change year to year — because the composer reads a dated claim as datable and a timeless one as a fact with no clock.
- 02Date the claim, do not decorate the sentence. A date that scoped nothing was lifted no more than a bare timeless one — the win comes from the date actually vouching for the fact being current.
- 03Put the date in the sentence. A "last updated" line the composer did not lift was invisible to the extraction — only the scope inside the sentence rides along into the card.
- 04Do not stamp stale. A visible old date was cited less than the same claim left undated, and on 7% of pages the composer preferred a competitor with a current date — refresh the stamp when you refresh the fact.
Where this argument breaks
For queries whose answer is a timeless fact — a definition, a physical constant, a mathematical relation that does not change year to year — there is nothing to date and the lever is irrelevant, so it is for answer sentences whose fact is time-sensitive: prices, limits, rankings behaviour, rates, version-specific claims. For facts a page cannot re-verify, stamping a date to satisfy this is the stale-stamping failure, not the win — an honestly-maintained timeless present beats a date you cannot stand behind. For navigational and brand queries there is no answer sentence whose shape matters. For narrative and historical passages — case studies, timelines, retrospectives — a past date is the content, not a freshness lever, and the temporal-scope pass is for the answer sentences on time-sensitive informational queries only. For some languages the effect may differ — in our parallel Chinese-language audit (文心一言, 元宝, 通义) the dated-claim win was present but the timeless present was carried more by a «目前» / «现阶段» frame the composer sometimes read as a soft current-scope and accepted, so the clockless-assertion problem was milder. The 2.3× figure on visibly-changing facts is the strongest cut and we are confident in it; the stale-stamp penalty is the one we would most want to re-measure as the composer's freshness handling changes. Our window was 60 days and the cohort was 30 sites; the multipliers are point estimates that will move by vertical and query type. Outside those carve-outs the lesson holds: in 2026 the AI Overview lifts an answer sentence scoped to a current time — the date stated in the sentence, kept as fresh as the fact is re-verified — more readily than one that leans on the plain timeless present, the unit is the individual answer sentence rather than the page, and the cheapest citation win on a time-sensitive query is to move the currency out of the masthead and into the sentence that makes the claim.