AI Max for Search Campaigns: What It Is and How It Changes Targeting
Explains what AI Max for Search is, how targeting changes, and how AI Overview and AI Mode placements fit, for advertisers and agencies evaluating automation.


A founder forwarded me a grader report last month: 78 out of 100. Green checks for Quality Score and account structure. One polite nudge about mobile bids.
That same account paid $184 per lead in a business that broke even at $110. In the previous 30 days, I found 312 search terms with spend and zero conversions, two campaigns capped by budget while branded traffic got chased, and conversion tracking that counted every pageview as a win. The grader saw none of it. It graded the paint while the engine burned oil.
I have run the three tools people usually mean when they type Google Ads grader: the old WordStream Performance Grader everyone still links to, WordStream’s current audit flow, and Adsgrader. They catch different surface problems, miss the same money problems, and leave you with a report instead of a fix. Here is what each does well, what it skips, and what I would fix first.
Most people do not run a grader because they love a score. They run one because spend rose 18% last month and nobody in the thread can explain why.
I have been that person at 1am, exporting search terms into a spreadsheet and finding one broad keyword ate $1,400 with three conversions to show for it. A free audit feels like a shortcut to that answer, without paying an agency $2,500 to tell you to pause it.
WordStream built the original version of that shortcut. Its free Google Ads Performance Grader runs an instant audit and scores four areas: Quality Score, account activity, impression share, and wasted spend. That makes it useful as a starting flag. It tells you where the paint is chipped.
It does not tell you whether the car makes money.
Every grader starts with the same easy-to-read API buckets. First is Quality Score, which Google breaks into three parts: expected CTR, ad relevance, and landing page experience. Each is rated Above average, Average, or Below average against other advertisers in the same auction.
Google is blunt about the limit: Quality Score is a diagnostic, not a KPI or an auction input. I used to tell clients to chase 8s and 9s. I was wrong. A 6 with tight intent and clean tracking beats a 9 that buys curiosity clicks.
The other buckets are wasted spend, impression share, and account activity:
That last one catches real neglect. If your last meaningful change was 47 days ago, you do not have management. You have rent.
Use the grade to decide where to look first, never to decide whether the account works.
WordStream’s Performance Grader remains the benchmark for third-party audits. Connect your account through OAuth and it calculates wasted spend from non-converting search terms, maps Quality Score distribution across active keywords, and flags whether changes were made in the last 30 or 90 days.
For an operator taking over an account spending $15,000 a month, that report can save 45 minutes of manual sorting. What it misses is everything downstream of the click. It cannot tell whether 42 conversions last week were genuine demo requests or spam form fills on a thank-you page.
A high score can coexist with an acquisition cost double your break-even.
Adsgrader is NP Digital’s free Google Ads performance evaluation tool. Use it as another quick audit perspective, not as a substitute for looking at conversion quality, margin, or sales outcomes.
That distinction matters because a clean account setup can still buy the wrong traffic. A report can identify settings to inspect. It cannot tell you whether those settings return cash.
An audit can flag account hygiene; it cannot validate profit.
Then there is Google’s native score: the Optimization Score in the Recommendations tab. Google presents it as a checklist for Google Ads best practices. I treat it as a platform checklist, not a business verdict.
It can recommend broad match keywords, higher campaign budgets, and auto-applied suggestions. Those may be sensible moves in some accounts. They may also increase spend before you have proved that the extra traffic meets your target CPA.
A grader measures adherence to rules, not whether those rules protect your margin.
A grader is a photo of an auction that never sits still. Say you spend $20k a month and run a report on Monday. It flags 12% wasted spend and a Quality Score dip in two ad groups. By Friday, a competitor raises bids 22% on your top three exact terms, search traffic skews mobile, and 40 new queries enter through close variants.
The auction moved. Your fix list is already old.
I used to send 30-page audits and watch half the findings expire before the client approved page four. The deeper limit is what the API never shows the grader: close rate, margin by product, and which calls turned into jobs. So it praises a high-CTR ad that brings price shoppers and flags a low-CTR ad that closes at 31% because the copy filters hard.
A static score grades account settings, not distance to profit. Read the audit as a triage list dated today, not a verdict on the account.
A grader hands you a list. The list rots.
Search terms pile up at 200 to 400 new queries a week on a $20k account. Budgets hit caps on Tuesday and starve Thursday. One broken tracking tag quietly turns 60 sales calls into zero recorded conversions. I learned to sort every audit into work that repeats and work that sticks:
If nobody owns the daily part, the monthly part never matters.
That is why I stopped assigning audits to junior managers during a Friday check-in. The work needs someone watching the auction when it moves, not when the calendar says to look.
groas runs that way by design: specialized models handle bids, budgets, search terms, and ad tests continuously. Every change is logged with reasoning, while a named strategist owns direction. I still want a human setting guardrails on CPA and spend. I do not want that human spending four hours pulling the same search-term report the grader gave you for free.
When a grader drops 40 warnings on your desk, do not start at the top of the PDF. Graders weigh missing assets and low impression share as equal crimes. They are not.
If an asset is missing, you lose a line of text. If conversion tracking double-counts or search terms bleed cash, you lose the business. Follow this order instead:
Run WordStream or Adsgrader when you want a quick audit perspective. Treat the result like a thermometer: it can tell you the account has a fever, but it will not write the prescription.
If you are tired of interpreting scorecards and want to see what is wasting your budget in real dollars, groas runs that diagnostic continuously and executes fixes inside your guardrails, 24 hours a day.