Google Ads can build promotion assets from offers visible on a website. The official automated-promotions help page describes eligible active Search and Performance Max campaigns with linked location assets and no manual promotions currently running. Sourced assets are labelled Google AI, can serve for less than 24 hours and may be re-enabled while the offer remains verifiable. Account-level controls allow opt-out. The document does not establish a universal activation date or guarantee campaign results.
For an Indian retailer, the interesting question is not whether the feature exists. It is which of your website offers Google could plausibly read, and whether each one is safe to advertise. That is an audit problem you can run yourself, on your own pages, before you touch a campaign setting.
What the automation actually reads
Inventory every public version of an offer: homepage banners, category strips, product badges and other visible page content. Keep the conditions understandable to customers, not just attractive in the headline. Do not assume that a PDF, image, pop-up or login-only offer is eligible, or that extraction will interpret it exactly as your team intends. Verify what actually appears in the account.
One important retail risk is inconsistent offer text across those pages. Your homepage banner, category page strip, product page badge and blog post may each describe a different version of the same sale. A crawler does not know which one your merchandising team considers canonical. It reads what is there.
Campaign settings still matter, but the first content audit is the offer text Google can see on your website. Automation can surface both accurate terms and stale or inconsistent wording.
A website-offer audit for Indian retailers
Run this as a page-by-page review, not a campaign review. For every live offer, record the following in one sheet.
- Offer text as written, copied exactly, including any threshold such as a minimum cart value.
- Validity window, with the end date and time visible in text, not only in a countdown script.
- Variant scope: does the offer apply to all sizes and colours, or only the ones featured?
- Stock position for the exact variant a customer will select, not the product-level total.
- Shipping terms: free-delivery thresholds, delivery estimates, and any pin-code exclusions.
- Page owner: the named person who can change or remove that text today.
Two failure patterns to check for are stale banners and threshold drift. A cached page or forgotten campaign strip may continue to show an expired coupon. A banner may state one minimum while the cart enforces another. These are audit scenarios, not a measured claim about how often advertisers encounter them.
Before you approve any offer for automated use, it is worth checking that a promotion is stock-ready, because an advertised offer creates an order expectation that your packing and delivery capacity has to meet.
Illustrative audit row: Page says “10% off selected blue notebooks through 31 October”; checkout applies the offer only to a different colour or adds an undisclosed minimum. Mark the offer hold, verify the product, threshold and expiry in the live checkout, then update the page and retest before allowing an ad asset to reflect it.
Manual asset or automation: a decision rule
You do not have to choose automation everywhere. A simple split can help a smaller catalogue.
| Condition | Reasonable choice |
|---|---|
| Offer is stable for weeks, wording is fixed, stock is deep | Let automation read it |
| Offer changes daily, or stock is thin | Keep a manual asset you control |
| Offer depends on a threshold or variant rule | Manual, until the page text states the rule plainly |
| Offer is seasonal and coordinated with operations | Manual, reviewed daily |
Google’s help page says automated promotions are bypassed for a campaign while a manual promotion is running on that campaign. This is a campaign-level running condition, not a blanket rule that no manual asset may exist anywhere in the account. If both modes are in use, review the specific campaign and timing before relying on automation.
Daily review and the removal illusion
Google's current help text says extracting promotion start and end dates is expected in a future change. A visible expiry date is still useful for customers and the merchandising team, but do not rely on it as a confirmed automatic stop rule. Review the served asset and account controls when an offer ends.
Because each automated promotion lasts under 24 hours and may be renewed while the offer remains on the site, the review cadence that matters is daily. Assign one person to check live offers against the audit sheet each morning: dates, thresholds, variants, stock, shipping.
A common misunderstanding deserves stating plainly. Removing a banner is not proof that a served asset instantly disappears. The documented behaviour is that assets are short-lived and renewable based on site content, which means the practical lever is the website text itself. Clean the source, then verify in the account rather than assuming propagation is immediate.
Consider a hypothetical example, labelled as hypothetical and not a description of any real programme. A seller lists a festive coupon on a category page with a minimum order value and a variant exclusion buried in a footnote. Automation reads the headline offer. A customer orders the excluded variant, expects the discount, and support spends the week explaining. Nothing in the campaign was broken. The page was ambiguous.
Controls, expectations and honest limits
Opt-out sits at account level in automated asset settings, so decide as a business, not per campaign, whether you want this class of asset at all. If you keep it on, pair it with the audit routine above.
Set expectations internally too. The official documentation does not promise conversions or ROAS, and no activation date is confirmed, so do not brief your team as though every account is enrolled on a fixed schedule. Treat the feature as available behaviour to verify in your own account.
Finally, keep the offer language boring. Plain text, one canonical version per offer, explicit dates, explicit thresholds, explicit variant scope. Boring pages are easier for humans to maintain and easier for automated systems to read correctly.
Conclusion
Automated promotions reward retailers whose websites say exactly what they mean. The feature reads visible content, creates short-lived assets tagged as AI, renews them while the offer stands, and can be switched off at account level. It does not guarantee results and is not tied to a confirmed rollout date. Your real work is upstream: audit every live offer for dates, thresholds, variants, stock and shipping, name an owner, and review daily. Decide deliberately between manual assets for volatile offers and automation for stable ones. Fix the page first, then let the system read it.
Image: cottonbro studio via Pexels. Illustrative retail sale signage; not a live Google Ads promotion or Indian shop.
