Google Ads Expert Roundup - September 2026

TL;DR: We asked three paid search practitioners the same questions about Google Ads. All three named AI and automation as the defining trend, and all three said the human edge is business context the algorithm doesn't have. Asked which change advertisers are underestimating, they gave three different answers: the August 17 bidding change, AI Max, and the upcoming AI Brief.

We asked three paid search practitioners the same set of questions this month.

The first one barely needed thinking about. AI. Automation. Google taking over more of the account.

Then we asked which recent Google Ads change advertisers are underestimating. Three people, three completely different answers.

That split is the useful part. The panel agrees on where Google Ads is going. They don't agree on which part of the hand-off is going to cost you.

• Julie Friedman Bacchini Neptune Moon & PPC Chat neptunemoon.comLinkedIn
• Meriem Nacer Paid Media Consultant, 4M Digital Consulting 4mdigitalconsulting.comLinkedIn
• Anthony McDaniel Sr. Paid Media Specialist, Divergent Digital divergent.digitalLinkedIn

Where the panel agreed, and where it split

Each panelist's answer to four of our questions, condensed. Full quotes below.

The most important trendAgreed

JulieAI, everywhere: in the results, in the backend, and in clients double-checking our work.

MeriemAI and automation taking more control of bidding, targeting and creative.

AnthonyAutomation, which makes clean conversion and audience data matter more.

What they see in accountsOverlapping

JulieCPCs keep going up, with less transparency into why.

MeriemSimpler structures, broader targeting, more reliance on automation.

AnthonyClients with consistent, quality signals keep growing.

The most underestimated changeSplit

JulieThe August 17 bidding change, which affects more than budget-limited campaigns.

MeriemAI Max, switched on without understanding what it opts you into.

AnthonyAI Brief, coming to AI Max and Performance Max.

Where humans still winAgreed

JulieBusiness data, test history and priorities Google doesn't have.

MeriemStrategy, context and judgement about what to optimise toward.

AnthonyData quality, when to restrict automation, and objective alignment.

The trend everybody named

Our first question: what's the most important trend shaping Google Ads today?

"The obvious answer is, of course, AI. Whether it is the ever expanding AI presence in search results or the application of AI in the backend or clients using AI to 'double check' everything we do. It is inescapable! Regardless of how you feel about AI (I hate it) we have to figure out how to coexist with it and where we can make it work for us and our clients."

"AI and automation, without question. Ad tools like Google, Microsoft and Meta are asking us to hand over more and more control, from bidding and targeting to creative. That can be incredibly powerful, but only when it has good data, clear objectives and the right guardrails.

Our role is shifting from pulling every individual lever to understanding what we're feeding the machine, questioning what it's doing and knowing when automation is actually helping the business, rather than just helping the Ad platforms spend more."

"The overall shift towards automation leading to the exponentially rising importance of clean, quality data/signals (especially conversion & audience data)."

Three very different temperatures. Julie hates it. Meriem calls it incredibly powerful, with conditions. Anthony skips the feelings and goes straight to the data.

Same conclusion, though. Nobody on this panel thinks the job is pulling levers anymore. The job is deciding what goes into the machine and questioning what comes out.

Meriem's last line is the one worth keeping: automation should be helping the business, "rather than just helping the Ad platforms spend more." Those aren't always the same thing.

What they're seeing in their accounts

Next we asked what trends they're seeing across their own clients, and whether those trends will continue.

"CPCs keep going up. This will continue for the foreseeable future, unfortunately. Between shrinking transparency into performance and the continued tweaking of the ad auctions, advertisers have little control over rising costs. Best we can do is have hard conversations about breaking points where the costs per click and by association the costs per acquisition make the platform a less attractive choice."

"I'm seeing simpler account structures, broader targeting and much more reliance on automation across clients. The biggest difference is that the quality of the data we feed the platforms matters more than ever.

I'm also seeing less value in over-segmenting everything just for the sake of control. We still need structure and guardrails, but we need to give the algorithms enough data and room to actually learn.

I absolutely expect that to continue. The challenge will be balancing automation with enough visibility and control to make sure it's driving real business results, not just better-looking platform metrics."

"Clients with consistent, clear, quality signals are seeing consistent growth, which I do expect to continue."

Read those together and they're one story told from three seats.

Accounts are getting simpler and broader, so more of the performance rides on the signals you send. The accounts with good signals keep growing. And everybody pays more per click, good signals or not.

Julie's "breaking point" deserves an actual number. At what CPC, or what CPA, does Google stop being the best place for this money? Work that out while you're still comfortably under it, not the month you cross it.

Three changes advertisers are underestimating

This is where the panel split. Each pick also sits at a different stage: one is already live in every eligible account, one is a setting you switch on, and one hasn't fully arrived yet.

Julie: the August 17 bidding change

"The August 17th bidding changes. It was announced and technically only impacts campaigns that were designated as 'limited by budget' but the reality is that changing the auction impacts everyone! Accounts that are not limited by budget should also be paying attention to their CPCs and CPAs and ROAS and adjust accordingly as this change continues to impact performance."

We made this change the headline of our August roundup. Since August 17, budget-limited campaigns on Target CPA or Target ROAS are optimized more closely toward the target you entered. A campaign that was comfortably beating its target now drifts toward it.

Julie takes it one step further, and it's the step most coverage skipped. The auction is shared. If budget-limited campaigns across your market start bidding closer to their targets, the auction you're in has changed, whether or not any of your campaigns are budget-limited.

So don't stop at the campaigns Google flagged. Look at CPC, CPA and ROAS on everything since mid-August.

Meriem: AI Max

"AI Max. I think a lot of advertisers are switching it on without fully understanding what they're opting into. It can expand where your ads show, change how your ads are put together and give Google significantly more control.

I'm not against it, but 'Google recommends it' isn't a strategy. Test it, understand what it's doing and make sure you have the right exclusions, data and guardrails in place before handing over that control."

AI Max for Search is one campaign setting that does three jobs. It matches your ads to searches beyond your keywords. It generates and customizes your ad text. And it can send people to a different landing page than the one you chose.

Each of those comes with controls, like brand and URL exclusions. Controls only help if somebody set them.

Meriem's test for whether you're ready is simple. If the only reason AI Max is on is that Google recommended it, you aren't.

Anthony: AI Brief

"It's an upcoming one, but the addition of AI Brief into AI Max and Pmax campaigns could be a real game changer as far as providing quality, contextual targeting & messaging information to those automated campaign types. It could be nothing, or it could be the next big lever we learn how to work with."

Google announced AI Brief on April 30. It's a plain-language brief inside AI Max with three parts: messaging guidelines (what your ads should and shouldn't say), matching guidelines (which searches to prioritize or avoid) and audience guidelines (what to say to whom). Google said it would start with AI Max for Search in English, then come to Performance Max and AI Max for Shopping.

Here's why Anthony might be onto something. Every panelist said the human edge is business context the algorithm doesn't have. AI Brief is the first input that asks for that context directly, in sentences.

If it works, writing a good brief becomes a PPC skill. If it doesn't, it's one more text box. Either way, draft yours before it shows up in your account.

Three underestimated changes, three different stages

How much an advertiser has to do before each change reaches their account, as of the end of September 2026.

1. Live. Nobody had to opt in

   The August 17 bidding change

   Flagged by Julie

   Check: CPC, CPA and ROAS since August 17 on every campaign, budget-limited or not.

2. Live. You switch it on

   AI Max for Search

   Flagged by Meriem

   Check: search terms, final URLs, and your brand and URL exclusions.

3. Rolling out. Search first

   AI Brief

   Flagged by Anthony

   Prep: what to say, what never to say, and which searches to chase or avoid.

Arrives on its ownNeeds you to actNot here yet

Where humans still make the difference

We asked: as Google gives its automation more control, where do human PPC experts still make the biggest difference?

"If all you do is let Google google, then you will not stand out in any way. Google has access to tons of data points, but they don't have access to tons of business specific data points that humans do! Nor does Google know what has been tested in the past, what the nuanced priorities are, etc. Only a human brain can do that."

"Strategy, context and judgement. Google can optimise towards the goal we give it, but it doesn't understand the wider business, the client's priorities or whether the data it's chasing is actually valuable.

That's where PPC experts make the biggest difference: deciding what we should be optimising towards, challenging what the platform is telling us, understanding the wider marketing mix and knowing when to intervene.

Automation can make decisions incredibly quickly. Our job is making sure it's making the right decisions in the first place."

"Confirming data quality, determining best uses of automation, and when to restrict automation, overall strategy/objective alignment decisions."

Not one of them said "writing better ads" or "finding keywords." Every answer is about knowing things Google can't: what the business actually needs, what's already been tried, which conversions are worth anything.

That knowledge only pays off once it reaches the account. It gets there through what you feed the algorithm, and through the moments you step in and say no.

What only you know has to travel through the inputs

Conceptual diagram built from the panel's answers. Left and right columns paraphrase the panelists.

What only you know

• Business-specific dataJulie
• What's been tested beforeJulie
• The client's real prioritiesJulie, Meriem
• Whether a conversion is worth chasingMeriem, Anthony
• The wider marketing mixMeriem

→

How it gets in

• The conversions you count and the values you pass
• Goals and bid targets
• Exclusions and guardrails
• Plain-language briefs (AI Brief)
• Stepping in: restricting or switching automation off

→

What the algorithm brings

• Tons of data pointsJulie
• Optimizing toward the goal you give itMeriem
• Decisions made incredibly quicklyMeriem

Feature deep dives

We also asked about two newer features. Julie and Anthony passed on these, so both takes are Meriem's.

Local customer optimization for Performance Max

This is a setting for PMax campaigns with store goals. It prioritizes people who are near a location and look ready to act, reading intent across Google Maps, Waze and local Search. It doesn't work with Merchant Center product feeds or online conversion goals.

"I like the direction of it, particularly for businesses where driving genuinely new local customers matters more than simply getting another conversion.

My caution is the same as with a lot of PMax features, it's only as good as the data behind it. If Google can't accurately identify who is an existing customer, you risk making optimisation decisions based on an incomplete picture.

So it's a useful feature, but I'd want strong first-party data, clear measurement and a proper test before relying heavily on it."

Her caution has teeth. Store-goal campaigns only support new-customer acquisition in "New customer only" mode, and that mode depends on Google knowing who your existing customers are. Give it a thin customer list and some of the "new" customers it finds will be people who've shopped with you for years.

Multi-campaign Search experiments

Google's budget and bidding target experiments for Search, rolling out in September, test one budget or ROI-target change across several Search campaigns in a single experiment, instead of one campaign at a time.

"Experiments aren't something I currently use much within Google Ads, so this isn't a feature I'm personally rushing to use. I can see the value of being able to test across multiple campaigns rather than being limited to one, particularly for accounts where individual campaigns have lower volumes. But for me, it would need to make testing genuinely easier and provide something I can't already learn from how I manage and test accounts daily."

The low-volume case is the real one. A single small campaign can run for a long time before an experiment gives you a readable answer, and pooling several campaigns into one test may be the only way some accounts get an answer at all. If your campaigns are big enough to test on their own, Meriem's bar stands: it has to tell you something your day-to-day testing doesn't.

What to do this week

1. Pull CPC, CPA and ROAS since August 17 for every campaign, not just the budget-limited ones.
2. Write down your breaking point: the CPC or CPA at which Google stops being worth it. Do it before you get there.
3. For every campaign with AI Max on, read the search terms and final URLs and check your brand and URL exclusions. If you can't say why it's on, turn it into a test.
4. Audit the conversions you optimize toward. If one of them isn't something the business would pay for, it shouldn't be steering bids.
5. Draft your AI Brief now: what to say, what never to say, which searches to chase and which to avoid.
6. If your Search campaigns are too small to test one at a time, try one budget or target experiment across several of them.

The panel didn't agree on what's most dangerous right now. They agreed completely on what still matters: the context Google doesn't have, and the judgement to know when its answer is wrong.

Thanks to Julie, Meriem and Anthony for the straight answers. Their work lives at Neptune Moon, 4M Digital Consulting and Divergent Digital.

And if CPCs keep climbing the way Julie expects, you want to catch the jump the week it happens, not in the month-end report. That's the job of anomaly detection.