Google Ads AI dashboards started showing up inside live advertiser accounts over the past week, following Google's original announcement back in August. If you manage a Google Ads account, you can now type a plain-language prompt and get a chart plus an AI-written explanation of why performance moved. Two more features, Local Customer Optimization and a store-sales feed through Data Manager, rolled out alongside it for multi-location businesses. None of this is optional. It's just there now.
At WebEpex we manage Google Ads and Performance Max spend for clients across GCC, Europe, USA, Canada, and Indian SMBs, and we started poking at all three features the week they landed in a live account.
What Actually Happened With Google Ads AI Dashboards#
Google first announced AI-powered dashboards, insight cards, and a Gemini-built benchmarking tool for Google Ads and Analytics back on 10 August, in beta, English-language accounts only. As of 8 September, those dashboards started actually appearing inside live accounts, not just the announcement blog post.
A paid search consultant named Thomas Eccel spotted it first and posted it to LinkedIn, which is usually how these things surface before Google confirms anything officially. It's part of a pattern we've been tracking all month. The same Gemini-summary push landed in Search Console a couple of weeks earlier.
The same week brought two more changes aimed at multi-location retailers, restaurants, and local service businesses. Here's where each one actually stands right now:
- AI dashboards and prompt-based reporting — live in some accounts since 8 September, rolling out gradually.
- Local Customer Optimization — a Performance Max toggle shifting budget toward shoppers nearby on Maps, Waze, and local Search. Live now for store-goal campaigns.
- Store sales in Data Manager — connects CRM or Google Sheets data to measure walk-ins. Coming in the following weeks, not live yet.
Does This Actually Change Anything for My Account?#
If you're not running Performance Max with a store-sales goal, or you don't have more than one location, no. Skip ahead. If you run PMax for a retail chain, restaurant group, or multi-branch service business, this is worth an afternoon this week, not a rebuild.
My read: the dashboard piece is the smaller story here. A chart you can summon with a sentence is convenient. A toggle that reroutes your Performance Max budget toward people standing near a location is the one that actually touches spend.
The Honest Trade-offs#
The AI summaries are genuinely useful for a first pass. You open the account, ask what happened last week, and get a plain-English guess at the cause in under ten seconds. That's real time saved on the boring part of the job.
Here's what I'd flag, and it's my opinion, not Google's: nobody outside Google has published how accurate that "why" explanation actually is. It's a language model narrating a chart, and the trade press is already raising the same question about attribution reliability across Google's newer AI ad tools generally. We treat the summary as a starting hypothesis. Never as the finding itself.
Local Customer Optimization has the opposite shape of problem. The mechanic is simple to explain, but it's a budget decision made inside an already-opaque campaign type, based on Google's own read of who's "nearby and in-market." For a five-location business that's a small bet. For a fifty-location chain spending real money on it every week, it's worth watching closely before trusting it.
How We're Handling It#
Our rule for any feature that shifts budget on its own, on Google or Meta, is the same: test it in a duplicated campaign first, never the live one, and give it a full week or two of real spend before comparing it against a control. We wrote about why after Meta's own AI ad connectors landed with a similar permissions catch earlier this month. That's how we're approaching Local Customer Optimization for the multi-location clients we run Performance Max for right now.
At WebEpex, the AI dashboard summaries specifically don't go into a client report without one of us checking the underlying numbers first. A Gemini-written paragraph is a shortcut to a hypothesis, not a report you hand someone.
What I'd Tell a Client Asking About This#
If you're single-location, this doesn't touch you yet, keep scrolling. If you're multi-location and running PMax store goals, duplicate the campaign, turn on Local Customer Optimization there, and compare cost-per-visit against your untouched control for seven to ten days before you decide anything, the same before-and-after discipline we push on page speed and conversion testing. Don't paste the AI dashboard summary into a board deck without pulling the raw numbers yourself. That's the whole list.
If you want a second pair of eyes on whether your account's actually set up to test this safely, send me what you're running and I'll tell you straight. Takes two minutes, costs you nothing. cal.com/webepex/growth-review
