Google's New AI Ads and Analytics Tools Are Gated: Check These Four Account Settings

Four settings decide whether any of this is yours

An agency lead reads the headline on Tuesday morning, opens the manager account she uses for eleven clients, and goes looking for the prompt box everyone is describing. It is not there. Nothing is broken. She is simply on the wrong side of a restriction that the announcement never printed and her account type never advertised.

Google's post of August 10, 2026, "Evolve your marketing with new AI tools," written by product management senior director Josh Moser, announces genuinely interesting things: AI Overviews at the top of the Google Analytics homepage, revamped AI-powered insights cards on the Google Ads homepage, Dashboards you build by typing a prompt, and a benchmark comparing your campaigns against anonymized averages from similar businesses. What no part of it does is tell you, in one place, whether you can open any of them.

The answer turns out to be four separate gates: your language, your account type, which of the two products you are in, and one admin toggle almost nobody has looked at. Three of those you cannot influence. One of them you can, today, in about ninety seconds. The reporting-side news read of this release is covered separately; this post is only about eligibility.

What Google labelled, and what it left blank

Before the gates, it is worth being exact about what the announcement actually asserts, because the labels are doing more work than the prose.

The three items carrying a footnote

The post uses exactly one footnote. It reads, in full, "Currently available in beta for English language accounts." Google attaches that marker to three items and only three: Ask Advisor, where the post says "That's why we introduced Ask Advisor"; the Google Ads homepage cards, described as "We've also revamped the Google Ads homepage to surface timely AI-powered insights cards personalized to your business"; and the new Dashboards.

Note the tense on the first one. The August post footnotes Ask Advisor's beta status; it does not begin it. That status was published at Google Marketing Live back in May, and what August adds is the same footnote stretched over two further surfaces. So the beta label you are reading this week is a May label with an August dateline on it, which changes nothing about eligibility but changes a lot about how urgently you should treat it.

The three items carrying nothing at all

The genuinely new items get no availability label whatsoever. Google writes that Analytics "now" has "new AI Overviews at the top of the homepage that give you an instant summary of important updates since you last logged in," that you can stay current by "opting into notifications via phone or email," and that it is introducing "a new benchmarking feature that compares your campaign performance against anonymized averages from similar businesses."

"Now" is not a status. Nowhere in the post does Google write "generally available." So the correct summary of the release is: three items in beta, one item explicitly coming soon, and three items with no stated availability at all. That is an unusual shape for a launch post, and it is the whole reason a checklist is more useful here than a feature list.

Item	Where	Label Google attached
Ask Advisor	Ads and Analytics	Beta, English-only
Insights cards	Google Ads homepage	Beta, English-only
Dashboards	Google Ads	Beta, English-only
Dashboards	Google Analytics	"Coming soon"
AI Overviews	Analytics homepage	None stated
Notification opt-in	Analytics	None stated
Campaign benchmarking	Ask Advisor	None stated

Gate one: the language on the account

The most-quoted line in the coverage is also the least examined, because "English-only" sounds like a footnote and behaves like a wall.

What the footnote actually restricts

"Currently available in beta for English language accounts" is a constraint on the account, not on the query you type. It is not a request to write your prompts in English. If the account is configured in another language, the surface does not appear at all.

For the many advertisers running Spanish, Portuguese, French or German accounts, that single clause removes all three footnoted items in one move. Not degraded, not partially translated. Absent. It is worth filing next to the target-based bidding change, which is the opposite kind of news: not optional, not gated, and it does not wait for a language rollout.

Why the Analytics wording is stricter than the Ads wording

Google's own documentation phrases the same restriction differently on each side, and the difference matters when you are auditing accounts. The Analytics help page states that "Ask Advisor is currently available in properties with English as the selected language," with plans to expand support to other languages.

Properties, not accounts. So on the Analytics side the gate is per-property, which means one organization can plausibly have an eligible property and an ineligible one sitting next to each other under the same login. If you are checking eligibility across a portfolio, check it per property rather than concluding once and generalizing.

Gate two: manager accounts are excluded

This is the gate no coverage of the August announcement carried, and for agencies it is the one that decides everything.

The line the announcement never printed

Google's Ads help page for Ask Advisor carries a sentence the blog post does not: "Note that Ask Advisor isn't currently available in MCC accounts." The same page states plainly, "Ask Advisor is in beta, so there may be limitations."

An MCC, or manager account, is how essentially every agency, freelancer and in-house team managing more than one brand actually works. It is the login you live in. Excluding it does not slightly inconvenience that group; it removes them from the product the entire announcement is organized around, because Ask Advisor is the surface the benchmark is reached through.

What this means if you manage accounts for other people

The workaround is unglamorous and it is the only one available. Two paths, and which one you are on is decided by portfolio size rather than by skill:

• Under about five accounts: log into each client account directly instead of working through the manager account, and check eligibility inside each one. Tedious, but finite.
• Above that: treat the whole release as unavailable for now and revisit when Google lifts the MCC restriction, because per-account logins stop being an audit and start being a job.

The broader lesson is procedural. A release note tells you what a company built. The help center tells you who is allowed to use it, and the two documents disagree about scope more often than anyone admits. When an announcement is thin on eligibility, the product documentation is where the real answer lives, and it is worth reading before the planning meeting rather than after it.

Gate three: the product, not the brand

The third gate is the easiest to misread because it hides inside a parenthetical, and parentheses are exactly what a skim-read drops.

Dashboards exist in one of the two products named

Google's sentence: the new Dashboards are "in Google Ads (coming soon in Google Analytics), you can transform raw data into clear visualizations with simple text prompts." Each report then generates a summary explaining the why behind the numbers.

Read the scope carefully, because it is easy to get backwards. The Google Ads half exists now and carries the beta footnote. The Google Analytics half is the one that is coming soon, with no date attached. One product, one place, today.

Why the direction of that parenthetical matters

If you invert it, you build a Q4 reporting plan around typing prompts into Analytics and getting charts back, and you find out in October that the feature you planned around was the half that had not shipped. It is a small misreading with a large blast radius, and it is the single most likely error a summary of this announcement will introduce.

Gate four: the admin switch under the benchmark

The first three gates are Google's to open. This one is yours, and it is the reason this post is a checklist rather than a news read.

What Google already documents about benchmarking

Benchmarking in Google Analytics is not a new idea invented on August 10. It is a documented feature with a published method. Google states that it "offers benchmarks in percentiles (median, 25th, and 75th) based on peer groups of businesses for a wide range of industry categories," available "across various standard reports in Google Analytics, including the Acquisition, Engagement, Monetization, and Retention reports, as well as on the Home page overview." The data "is refreshed every 24 hours."

Your peer group is not something you pick from scratch either. Google says the "industry category is determined by a combination of factors including the broad industry category provided in setup and signals from things like a property" and its URLs and app attributes, and that "you can select sub categories within each of the top level categories listed above," down to a level as specific as athletic shoes.

The prerequisite most accounts have never opened

Here is the actionable line, and it is unambiguous: "To be eligible for benchmarking data in Google Analytics, your property must have the Modeling contributions" and business insights setting enabled, in the admin account settings. Enabling it means sharing aggregated, de-identified data with Google and others, which is a real decision with a real privacy trade, not a formality.

Whether the new Ask Advisor campaign benchmark inherits that same prerequisite, Google does not say. The feature is introduced and then simply left there, with nothing said about timing, qualification, or which industries get a peer group first. Search Engine Journal was the only outlet to press on that silence in its same-day coverage, and it came away with the same nothing.

Which is exactly why this is worth doing anyway. It is the only gate of the four you can act on, it takes a minute, and if the new benchmark does inherit the condition, an account that already shares data is an account that qualifies on day one instead of discovering the toggle a month late.

The floor underneath all of it

There is one more condition, and it disqualifies quietly rather than loudly. Google states that "Properties must also have a minimum volume of users and be generating a minimum volume of meaningful data in order to be included in a peer group and contribute to its benchmarking metrics," and that thresholds ensure a minimum number of properties exist before benchmarks appear for a peer group at all.

Translated: a small account can pass every other gate, switch on every setting, and still see nothing, because peer benchmarking is a feature that gets better the bigger you already are. That is a reasonable privacy design. It is also worth knowing before you build a quarter's reporting story on a comparison that may never populate.

Stacked up, the documented benchmarking gate is three conditions rather than one:

• The admin toggle. Modeling contributions and business insights, switched on in account settings.
• A resolvable peer group. Derived from your setup category plus signals from your URLs and app attributes, refinable down to a subcategory.
• Enough traffic. Both from your property and from enough peer properties for the group to exist at all.

The four gates, applied to one real account

Take a case that is common rather than convenient. A skincare brand doing about $40,000 a month, Spanish-language Google Ads account, Analytics property set to Spanish, campaigns run by a two-person agency out of a manager account.

Run the checklist:

• Language: the account is not English. All three footnoted items are out.
• Account type: the work happens in an MCC. Ask Advisor is out, which takes the benchmark with it.
• Product: prompt-built Dashboards are a Google Ads feature. Nothing changes in their Analytics property.
• Admin toggle: worth enabling regardless, but with the first two gates closed it changes nothing this quarter.

Four gates, zero opened. That is not a pessimistic reading; it is arithmetic, and it applies to a very large share of accounts outside the United States. The honest planning conclusion is that for this brand the August 10 release is a roadmap item, and the quarter still has to be won with something else.

What this release does not touch

Which brings up the part our readers care about most, and the part no headline mentioned.

The verb inventory

Take the announcement's own verbs and lay them next to each other. Summarize. Surface. Visualize. Compare. Every one of them is something you do to numbers that already exist, and not one of them is something you do with a camera. Google states the ambition plainly enough at the end: "Built with Gemini, these new solutions will help marketers move from insight to action even faster." Read that sentence for its subject rather than its adverb. Google supplies the insight; you are still the one taking the action.

Search the body of the post for the vocabulary that would signal otherwise and the result is a clean sweep of zeroes:

• creative: does not appear.
• asset: does not appear.
• Performance Max: does not appear.
• Demand Gen: does not appear.
• Merchant Center: does not appear.

That is a measurement release, end to end. Whatever is gated for you here, none of it was going to touch the ads themselves.

Where creative generation actually lives

This needs stating carefully, because the convenient version of it would be wrong. Google Ads generates creative today, in a different part of the product, documented separately: the help center puts it as "Google Ads offers generative AI tools to help you create new image assets in seconds." Around that sit published expectations for labeling AI-created or AI-edited assets and a separate policy governing AI-generated images. All real, all shipping, and all untouched by the August 10 post.

So the useful conclusion for anyone who makes video ads is narrow and worth keeping narrow. One announcement, on one date, did nothing to the production side. Two advertisers reading the same peer median will still be separated by what they can put in front of it, which is the argument creative analytics has spent years making from the measurement direction and this release accidentally makes from the other.

How Novoads solves the half that stayed manual

Nothing in the August release is gated for the brand above because of anything they did wrong. Their language, their account type and their product are all fixed. What is not fixed is how many ads they can test this month, and that is the half Novoads is built for.

Nothing here waits for a footnote

The inputs are a product image and a script, and the output is a vertical or horizontal ad you can upload the same afternoon. What matters for this particular post is what is absent from that sentence: no account language, no manager-account exclusion, no admin toggle, no beta list to be on.

The dial you turn between renders is deliberately small, because that is what makes a result readable:

• The hook. Hold everything else still and rewrite the first three seconds.
• The actor. Same script, different face, same product.
• The format. Vertical for Shorts and Reels, horizontal for Display.
• The length. Eight seconds for a scroll-stopper, a full minute for a demo.

Realism is not the argument, because every serious tool in this category rents the same frontier engines, Google Veo 3.1 included. The argument is what it costs to reach an ad you can actually run, and you know the loop is working when the variants start disagreeing with each other in the results. That is all an ad creative testing programme ever is.

What one more variant costs

Take the Pro plan at $69 a month for 100 credits. That values a credit at $0.69, a standard eight-second UGC ad at 4 credits or about $2.76, and a supporting product photo at roughly $0.21. Eight hooks against one product therefore costs 32 credits, near $22, or about a third of the month, to answer the question a peer benchmark can only pose.

Cost scales with the engine rather than with a plan tier, so a five-second clip lands near $2.07, a one-minute talking-actor ad near $6.90, and the top-end options nearer $11. A video has no single price, which is why the comparison worth running is against what UGC creators charge rather than against any headline figure. $1 for 3 days of access. Cancel anytime.

The gates you can open are the ones worth checking

Google shipped a good measurement release and labelled it carefully, if you read the labels. Three items are in beta for English-language accounts, one is explicitly coming soon in one of the two products named, three carry no availability label at all, and one restriction that removes most agencies from the whole thing lives in a help page rather than the announcement.

Check the four gates before you plan around any of it, and open the one that is yours. Then spend the rest of the week on the half of advertising nobody gated: how many different ads you can put in front of the audience that Google is now so good at describing.