When to test AI Max, which settings need attention and how to judge results through sales and lead quality.
Start with the business decision
AI Max is worth discussing once the business knows what the campaign needs to improve. More enquiries can create extra work without extra contracts. A hotel might gain bookings for its cheapest stays while earning little additional margin. Choose a primary outcome before changing settings: the cost of acquiring a valuable customer, revenue from confirmed bookings or profitable sales. Agree how much deterioration is acceptable while the test runs.
What AI Max changes
AI Max is a set of features for Search campaigns. It expands search term matching and automates assets, including text and landing page selection. It does not transfer responsibility for the campaign to Google. The business still owns its offer, conversion data and operating limits. Enabling a feature will not repair measurement that gives an irrelevant enquiry the same significance as a signed contract.
Define the boundaries
Review search term matching, text customization, final URL expansion and brand settings. Decide which pages may receive paid traffic. An expired package, careers page or terms page should not accidentally replace a sales offer. Check generated copy for promises, service scope and current prices. Apply the available exclusions deliberately rather than assuming the default configuration reflects the way the business makes money.
Compare the options in an experiment
Google provides AI Max experiments. Check campaign eligibility, control and treatment settings, and the primary success measure before starting. Brand settings added during setup can affect both arms. Avoid changing the offer, budget, website and conversion definition at the same time if you later want to attribute a difference to AI Max. Record the hypothesis and stopping criteria before results arrive.
A hotel example: dates matter
Suppose a hotel needs weekday bookings but already sells weekends well. A campaign that brings more weekend reservations may look successful in the platform while doing little for the actual objective. Examine stay dates, cancellations, booking value and acquisition cost. This is an analytical example, not a claim that AI Max targets according to hotel occupancy. Reservation data must help the team evaluate the outcome.
Complete the assessment in sales
Compare cost, sales value and customer mix after the test. Allow for the delay between click and purchase: recent leads may not yet have an outcome in the CRM. Check whether an apparent improvement came from a larger share of easy searches for the company's own brand. With few transactions, several days cannot support a strong conclusion. Insufficient evidence is a valid finding; a winner is not compulsory.
What the specialist still does
The specialist sets the objective, checks data quality, defines automation limits and interprets the result in the business context. They should explain why the test is needed and which decision it will support. Begin with a settings record, a baseline and one hypothesis. Expand only after evaluating sales. Automation can reduce some operational work, but accountability for the budget and outcome remains with the team.
Related articles
- Google Ads AI Max reporting: search query, ad and landing page
- Attribution versus incrementality: did the ad create extra sales?
- Google Ads for hotels: how to run campaigns that sell stays
FAQ
Does AI Max replace Search campaigns?
It is a set of features within Search campaigns. Evaluate the degree of automation through a controlled test tied to a business objective.
When should a test wait?
When measurement is unreliable, lead quality is unknown or the offer is changing substantially. Establish a usable baseline first.
Sources
Information checked: 25 September 2026. Feature access may depend on the account.