AI-Generated Product Titles Now Have a Google Ads Report: The Feed Guardrails Playbook

Google is quietly rolling out a new Product titles report inside Google Ads, and it is the first tool that shows advertisers exactly which AI-generated product titles Google is serving in place of the ones they wrote. The report entered public beta this week and was first spotted in accounts on August 26, 2026. If you run Shopping or Performance Max, this is the visibility layer you have wanted since Google started rewriting your titles at match time (Search Engine Roundtable).

The report tells you a story your Search terms report never could. Google’s model has been swapping your original titles for AI-generated variants whenever it thinks the swap wins the auction, and until now you had no way to see the swaps or measure them. That changes today.

What the AI-generated product titles report actually shows

The report is a table inside Google Ads that lists a sample of product titles Google AI created and served in your ads. Each row shows the AI-generated title alongside your original title, plus the standard performance columns: impressions, product clicks, click-through rate, cost, and average cost per click. In other words, it is an A/B result you did not know you were running.

Google is clear about when the AI titles appear. The banner inside the report reads that your original title will still be shown when it is the most relevant, and the AI title only enters the report once it has received enough impressions. That means the report is a lagging view. If Google served an AI title one thousand times last week, you will see it. If Google served it fifteen times, you will not.

The AI titles pull from the same catalog data as your original titles, but Google’s model can rewrite anything: word order, capitalization, attribute placement, and even which attributes make the 70-character cutoff that Shopping shows in standard results. AI-generated Shopping titles must be submitted through the structured_title attribute with digital_source_type set to trained_algorithmic_media (Google Merchant Center Help). What the new report shows is not your submitted AI titles. It shows the titles Google itself generated when it thought yours could be beaten.

Why the AI-generated product titles report matters this quarter

Three reasons. First, you can finally quantify the swap. If Google’s AI title is beating yours by a wide CTR margin, that is a signal your original title is under-optimized. If your title is winning, that is proof your feed team’s work is doing what it is supposed to do.

Second, the report exposes a testing pattern most advertisers did not know they were part of. AI titles enter the report only after enough impressions, so what you are seeing is Google’s own confidence that a variant is worth showing at scale. Read the swap list as a curated best-of, not as a comprehensive audit.

Third, the timing lines up with the January 31, 2027 image size deadline that already forces every ecommerce team to touch the feed. If you are auditing images anyway, adding a title audit costs almost nothing on top. Our Merchant Center image size deadline playbook covers the image half of that combined audit. This article covers the title half.

Your first Product titles report review this week

The report is beta, so the first read matters more than usual. Give it a proper session before it becomes the tab you glance at once a month and forget. Five moves get you a clean baseline.

One, open the Product titles report and export the full table. Do not read it inside Google Ads. Export to a spreadsheet so you can sort, filter, and share. The Google Ads UI hides the pattern.

Two, sort by cost descending. The AI titles that spent the most money are the ones you fix or accept first. A ten-dollar swap in a report that spent five hundred thousand dollars this quarter is not a priority. A two-thousand-dollar swap on your top SKU is.

Three, calculate CTR delta per row. The columns give you what you need. If the AI CTR is more than 15 percent above your original title CTR, Google’s model is signaling a real fix. If the AI CTR is within 5 percent, the swap is noise and your title is holding up.

Four, tag every row with a category. Winner original, winner AI, or noise. This tagging is the foundation of every action you take next. Do it inside the spreadsheet. Do not skip it.

Five, pull the AI-generated product titles column into a second sheet and look at the patterns. Are the wins coming from word order changes? Attribute reordering? Different key attribute placement? The pattern is your rewriting template for the rest of the catalog.

Rewrite your feed titles based on what the report shows

The report is only useful if it changes your feed. Sitting on the data without acting on it just tells Google to keep rewriting. Three moves turn the data into a feed change.

First, rewrite the winning AI titles into your primary title attribute directly. When Google’s AI title beat yours by a wide CTR margin, take that title, adapt it to fit your brand voice, and push it into the title attribute for that SKU. Do not overthink the copy. If Google’s model beat you by 30 percent, its structure is the one to copy. Our Content API migration playbook covers the feed plumbing you need in place to push title changes at scale.

Second, decide whether you want to submit your own AI titles through structured_title. If your team already uses a generative tool to draft titles, submit them through structured_title with the correct digital_source_type flag. This is not the same field as the AI titles Google generates. It is your controlled AI submission. The two coexist in the same account. If your team is spinning up AI-driven creative work across the feed, our AI image generation tools for ads guide covers the disapproval-safe workflow that keeps AI creative from tripping the same policy checks your titles do.

Third, add text guidelines in your campaign settings. Google’s beta report explicitly notes that advertisers can add text guidelines to steer AI-generated ad copy toward brand voice. If you have not set them, the AI has no guardrails to follow. The guidelines are the closest thing Google has offered to a per-brand style control.

Where AI-generated product titles quietly hurt performance

Not every AI title win is a real win. Two failure patterns cost accounts money, and both are invisible without the new report.

The first is the CTR-conversion divergence pattern. Google’s AI title can lift CTR by rewriting toward a more clickable variant while dropping conversion rate on the landing page. The report shows CTR but does not show conversions or revenue per click. You have to pull those separately from your standard reports and cross-reference by product ID. If CTR is up 20 percent and conversion rate is down 15 percent, the AI title is a loss in disguise. Our AI Max for Shopping playbook covers the retail bidding stack this pattern most often shows up in.

The second is the brand voice drift pattern. Google’s AI does not know your voice. It knows what wins the auction. That is why the model may drop your brand name from the front of the title in favor of a more descriptive lead. In categories where brand recognition drives conversion, this rewrite costs you real revenue even when the CTR looks fine. Text guidelines help, but they are not a hard filter.

Both patterns are why you tag every row in step four of the review. Winners without conversion data are provisional. Only the winners that hold up after you cross-reference revenue belong in your rewrite queue.

Fit the AI-generated product titles report into your monthly cadence

The report is beta today, but Google’s product pattern is to graduate high-visibility reports into standard columns within a quarter or two. Set up your workflow now so you are not scrambling when it goes GA.

Add the report review to your monthly account audit. The first read this week is a baseline. Every month after, pull the report on the same day and diff the swap list against last month. New AI titles entering the report are new signals to act on. Old AI titles falling out mean either the swap did not hold up in the auction or you already fixed the original title in the feed. If your setup runs Shopping alongside a Performance Max stack, our Performance Max hybrid strategy guide covers how title rewrites propagate across the ecommerce campaign layer.

If your account runs across many brand feeds or many client accounts, treat this as a per-account review, not an aggregate one. The AI title patterns that win in one vertical do not carry over to another. Our writeup on the Merchant Center for agencies setup covers the account architecture that makes per-account audits sustainable.

Google shipped this report to give you visibility, and the accounts that read it will start seeing feed wins their competitors are still leaving on the table. If your team wants a partner to run the first review and turn the swap list into a real feed rewrite plan, book a walk-through on our free consultation page. Our media buyers have shipped this exact audit for retail accounts since AI titles first started serving. Let’s Grow!