Google AI Mode Shopping Ads: What to Change (2026)
Open your Google Ads account and look for where Google AI Mode shopping ads show up. Go on, look. There's no campaign type for it, no toggle, and no line in your placement report. Yet Google says AI Mode has passed 1 billion monthly users, and ads now sit inside those conversations.
That's the awkward part of this change. You can't buy the spot, you can't block it, and you don't write the ad. Gemini does. What you do control is everything underneath: your Merchant Center feed, the campaigns you run, and how you read the results.
One quick note on numbers. You'll still see "75 million daily users" quoted in articles about AI Mode ads. That figure dates from Google's Q3 2025 earnings, and it's the most recent daily number we could find. The newer one is 1 billion+ monthly users, shared at Google I/O in May 2026.
Quick answer: what should Google Ads advertisers change?
Google AI Mode shopping ads are generated by Gemini from your Merchant Center feed, so advertisers should change three things: enrich product data (titles, descriptions, conversational attributes), run campaign types that are eligible for AI placements (Shopping, Performance Max, AI Max), and measure with proxy metrics, because there's no dedicated AI Mode report yet.
The rest of this guide shows how, ending with a 30-day plan that finishes before Black Friday (November 27, 2026).
What changed in Google Ads and AI Mode in 2026?
AI Mode went from a test to a real ad surface in under a year. Here's the timeline, newest first. Where a date comes from a single trade source rather than Google, the table says so.
| Date | What happened | Source |
|---|---|---|
| Oct 12, 2026 | Offer badges reported in AI Mode Shopping ads for brands with approved Merchant Center promotions | eCommerce Fastlane, citing Common Thread Collective |
| Aug 31, 2026 | Shopping ads reported in carousel format inside AI Mode answers, served through existing Shopping and Performance Max campaigns | Common Thread Collective |
| May 20, 2026 | Google Marketing Live: AI-powered Shopping ads, Conversational Discovery ads, Highlighted Answers, Business Agent for Leads, expanded Direct Offers, six conversational attributes in Merchant Center | Search Engine Roundtable, SEJ |
| May 19, 2026 | Google says AI Mode has passed 1 billion monthly users | SEJ |
| Apr 30, 2026 | AI Max for Shopping launches: text customization, Final URL Expansion, Optimal Format Selection | Google Ads blog |
| Feb 11, 2026 | Google reported to be testing Shopping ads and Direct Offers inside AI Mode | eMarketer |
| Oct 2025 | Google's Q3 2025 earnings: AI Mode at 75 million+ daily users, the daily figure still quoted today | Presence AI |
One caution. Coverage doesn't fully agree on where each new format runs (inside AI Mode, or on regular Search results), and several formats are still US-first or in testing. Check what your own account and your own search results actually show before you rebuild anything.
How do shopping ads work in Google AI Mode?
A shopper asks a long, messy question. Gemini answers it, and a sponsored product can appear inside or beside that answer, clearly labelled. Google's own Marketing Live announcement describes AI-powered Shopping ads, where Gemini picks your most relevant product and writes a short explainer for why it fits that exact question.
Here's what you can say with reasonable confidence:
- There's no AI Mode campaign type. Eligibility comes from running Google's AI-powered setups: Performance Max, AI Max for Search and Shopping, Shopping campaigns, and broad match as Dynamic Search Ads move into AI Max. Smart Bidding is part of the requirement (SEJ).
- You can't opt out. The same SEJ guide notes advertisers can't block their ads from appearing in AI Mode.
- Billing looks familiar. Common Thread Collective reports these placements run on cost-per-click through your existing campaigns, with no premium announced so far (source).
- Queries are longer. Google says AI Mode searches run about three times longer than classic ones, so a one-line product title has less to match against.
The practical difference from classic Shopping is easiest to see side by side:
| Classic Shopping ads | AI Mode shopping ads | |
|---|---|---|
| What triggers the ad | A keyword match | The whole conversation plus your feed data |
| Who writes the words | You, through feed titles | Gemini, from your feed |
| Your main lever | Bids, negatives, product groups | Feed quality and campaign structure |
| Reporting | Search terms and placement reports | No dedicated AI Mode report yet |
What are the 7 changes Google Ads advertisers should make?
In order of impact, from the fixes that cost nothing to the ones that need a test.
- Fix Merchant Center before you touch a campaign. AI Mode ads are built from your feed, so thin titles and one-line descriptions produce thin ads. Open Diagnostics, clear disapprovals and missing attributes, and start with your highest-spend categories and best sellers, not the whole catalogue.
- Write titles and descriptions for questions, not keywords. Nobody types "trail shoes" into a conversation. They type something like "waterproof trail shoes for wide feet under $150." Put the details shoppers filter on (material, fit, size, use case) at the front of the title, and give descriptions of real substance.
- Add conversational attributes to your top products. Google launched six optional Merchant Center fields in May 2026 to help AI surfaces understand products. They don't affect product approval, and Google recommends sending them through a supplemental data source. Details in the next section.
- Check your campaign structure for eligibility. If you run standard Shopping, AI Max for Shopping is worth a controlled test. Google calls it a one-click upgrade that keeps your product targeting and bidding controls, and you can switch Final URL Expansion off (Google Ads blog). Record your baseline first and run it as a test, not a blanket switch. You may read that AI Max will replace Performance Max for Shopping. Google's own April post says Performance Max remains the choice for cross-channel performance, so verify before restructuring on a rumour.
- Don't strangle delivery with tight targets. Common Thread Collective warns that an overly aggressive Target ROAS or CPA can keep products out of auctions entirely. That's worth checking: look at impression shares on your best sellers before and after any target change.
- Get promotions approved early. AI Mode now offers badges for products with approved Merchant Center promotions, per eCommerce Fastlane. Its guidance is to submit Black Friday promotions by about November 20, a week before November 27, test every code at checkout first, and attach promotion assets in Performance Max too.
- Reset expectations upstairs. There's no clean AI Mode report, so tell your boss or client this is positioning for where search is heading, not a line item you'll optimise weekly. Better they hear it from you now than discover it in a review.
Which feed fixes decide whether Gemini picks your product?
Think of the feed as the ad itself. Gemini reads it the way a shopper reads a product page, so the question is whether it has enough to work with. Here's an illustrative before-and-after (a made-up product, to show the pattern):
| Field | Thin version | Gemini-friendly version |
|---|---|---|
| Title | Blue Shirt | Men's Slim Fit Oxford Shirt, Blue, 100% Cotton |
| Description | High quality for every need. | Brushed cotton oxford with a slim cut that sits close to the body. Size up for a relaxed fit. Not designed for formal black-tie dress codes. |
| Question and answer | (empty) | Is it machine washable? Yes. Cold wash, tumble dry low. |
The description on the right does something the left one can't: it says who the product is for and who it isn't for. That's exactly what a conversational model needs to say "this one fits your question" or "skip this one."
Then add the six conversational attributes Google introduced at Marketing Live. All are optional, and Google says to avoid repeating details already in your description, highlights or product details.
| Attribute | What it helps answer | Where to mine it |
|---|---|---|
| question_and_answer | Pre-sale questions | Support tickets, presale chat, product page FAQs |
| document_link | Spec and sizing questions | PDF manuals and size guides |
| related_product | What goes with this? | Your catalogue: accessories, required parts, substitutes |
| item_group_title | Which family a variant belongs to | One shared name across all variants |
| variant_option | How the variants differ | Size, width, material, colour |
| popularity_rank | How well it sells in your range | Your own sales data, scored 0 to 100 |
Send them through a supplemental data source so an app sync can't overwrite them, and pilot on 20 to 50 best sellers before scaling. Google's own help page lists the exact formats and limits for each field, and Search Engine Roundtable's write-up is a quick overview.
Recommended resource
Want your Merchant Center feed audited before Black Friday? Book a free Google Ads and Merchant Center audit.
Book a Free Feed Audit
How do you measure AI Mode shopping ads with no dedicated report?
Honestly, partly you can't. Search Engine Journal's August guide put it bluntly: there's no clean AI Mode view for ads today. So use proxies, and be upfront that they're proxies.
- Annotate your account. Note the dates in the timeline above in Google Ads and GA4, so a shift in impression share isn't mistaken for a campaign problem.
- Watch impression share on best sellers. It's the closest signal you have for whether products are entering more auctions.
- Read AI Max search term data. It shows the longer, conversational queries reaching your campaigns, which is useful language to feed back into titles and descriptions.
- Track branded search. It's a lagging sign that discovery is working.
- Open AI Performance Insights in Merchant Center. It reports your share of voice across organic AI Mode and AI Overviews results. It's organic only, English-language, and still rolling out by market, so check whether your account has it.
- Look for a placement breakdown. Some agencies expected AI Mode to appear as its own segment in reporting. At the time of writing we couldn't confirm one has shipped.
What mistakes should advertisers avoid?
- Looking for an AI Mode switch. There isn't one. Eligibility comes from your campaign setup and feed.
- Switching AI Max on everywhere at once. One-click upgrades still change how ads are matched. Test one campaign, compare, then expand.
- Promising prices or dates in product Q&A. Google's guidance keeps time-bound details out of those fields. Put offers in promotions instead.
- Treating the 75 million figure as today's reach. It's a Q3 2025 daily number. Use the 1 billion monthly figure and say which one you mean.
- Waiting for a report before acting. With Black Friday on November 27, the feed work is the part you can do now.
What does a 30-day AI Mode action plan look like?
This plan runs from today, October 21, to the Merchant Center promotion deadline of about November 20, just ahead of Black Friday on November 27, 2026.
| When | What to do | Owner | Done when |
|---|---|---|---|
| Oct 21 to 27 | Audit Merchant Center diagnostics, fix disapprovals, and record baseline impression share, CTR and ROAS for Shopping and Performance Max | Paid media | Baseline saved and top-50 SKU list agreed |
| Oct 28 to Nov 3 | Rewrite titles and descriptions for the top 50 SKUs so they answer real shopper questions | Merchandising and paid media | Updated feed has processed in Merchant Center |
| Nov 4 to 10 | Build a supplemental feed with question_and_answer, variant_option and related_product for 20 to 50 best sellers | Merchandising | Values show correctly on the processed products |
| Nov 11 to 20 | Test AI Max for Shopping on one campaign, and submit and test every BFCM promotion in Merchant Center | Paid media | Test is live with a baseline, and promotions are approved |
Frequently asked questions
Tap a question to open the answer.
Related reading
For the wider picture of how paid, organic and AI search fit together, read our guide to AEO vs SEO vs Google Ads for US ecommerce. If you also sell on Amazon, see how Rufus and ChatGPT Shopping change listings, and before the rush, run our Shopify holiday readiness audit.
Need help getting your account ready for AI Mode?
Cotasks is a Google Partner agency working across the UAE, USA, UK, Saudi Arabia, Australia and India. If you'd rather not untangle feeds, campaign structure and reporting alone before Q4, our team can:
- Audit your Merchant Center feed and fix the gaps that hold back AI placements
- Rewrite titles and descriptions for your top products, and build the supplemental feed for conversational attributes
- Restructure Shopping, Performance Max and AI Max campaigns, and set up a clean baseline and test plan
- Set up reporting that's honest about what AI Mode can and can't tell you
Book a free Google Ads and Merchant Center audit with Cotasks, and we'll show you where your feed stands before the holiday rush.
Recommended resource
Book a free consultation to get your account ready for AI Mode.
Book Free ConsultationWritten by Rahul Bisht, digital marketing at Cotasks.
Sources
Pages opened for this article (all checked October 21, 2026):
- Google Ads blog: Adapt your Shopping campaigns to modern Search with AI Max, April 30, 2026
- Search Engine Roundtable: Google AI Mode Shopping Ads, Business Agent for Leads & Direct Offers, May 21, 2026
- Search Engine Journal: How to Advertise in Google AI Mode for Ecommerce, August 26, 2026
- Common Thread Collective: Google AI Mode Now Shows Shopping Ads Inline, August 31, 2026
- eCommerce Fastlane: Google AI Mode Promotions for BFCM 2026, October 14, 2026
Also used from search results (not opened in full): Presenc AI usage statistics, eMarketer on AI Mode shopping ads and Search Engine Roundtable on conversational attributes.