Google describes AI Max as a suite of targeting and creative features for existing Search campaigns. It is not a replacement campaign type. Enabling it can turn on search-term matching, while asset optimisation includes text customisation and final URL expansion controls.
That distinction matters. “Turn on AI” is not one reversible setting with one predictable effect. Advertisers need to understand which controls change, what data the system will use, and how success will be measured.
What AI Max can change
According to Google Ads Help, AI Max includes two main areas.
Search-term matching
This can use broad match and keywordless technology to find queries beyond the account's existing keywords. Google says the system learns from current keywords, creative assets and URLs.
The opportunity is incremental relevant demand. The risk is spending on queries that look statistically related but do not match the business's actual service, margin, location or qualification rules.
Asset optimisation
Text customisation can adapt ad copy, while final URL expansion can select a different landing page from the domain when Google predicts it is more relevant. Those features increase the importance of website quality. An outdated, thin or unsuitable page should not be eligible to become an automated destination.
Preconditions before testing
Reliable conversion data
Optimisation is only as useful as the outcome signal. Separate qualified leads from spam, duplicate submissions and low-value actions. Where sales close offline, import verified outcomes when the consent and technical setup allow it.
Landing-page governance
Review indexable and advertising-eligible URLs. Exclude pages that are outdated, legally sensitive, irrelevant to the campaign or unable to convert. Confirm that location and service claims are accurate.
Budget capacity
Google notes that AI Max will not be effective when a campaign is limited by budget. That is a platform statement, not a reason to increase spend automatically. It means the test design must account for whether budget constraints prevent a fair comparison.
Brand and location controls
Document brand inclusions or exclusions, locations of interest, negative keywords and campaign-level restrictions before launch. Save the baseline so any setting change can be traced.
A defensible experiment design
Google provides AI Max experiments that split traffic between a control with AI Max off and a treatment with selected settings enabled.
Use a written test plan:
| Decision | Define before launch |
|---|---|
| Hypothesis | Which incremental demand or efficiency gain is expected? |
| Primary outcome | Qualified leads, revenue, profit or another verified conversion |
| Guardrail | Maximum acceptable cost, irrelevant-query rate or brand risk |
| Scope | Campaigns, locations, devices and settings included |
| Duration | Long enough to cover conversion delay and normal weekly variation |
| Review | Search terms, selected URLs, generated assets and lead quality |
Do not change bids, budgets, offers, tracking and landing pages repeatedly during the same experiment unless the design accounts for those changes.
What to inspect after launch
Review more than aggregate conversions:
- search terms and the matching route that triggered the ad;
- landing pages selected through final URL expansion;
- generated or adapted asset language;
- new versus existing customer quality;
- cost and conversion delay;
- geographic accuracy; and
- offline sales acceptance or rejection reasons.
Google's reporting documentation describes search-term views with headline and URL information as well as asset reporting. Use these to identify where automation adds value and where exclusions or page improvements are needed.
What AI Max does not fix
It does not repair an unclear offer, weak conversion tracking, poor lead handling, unsuitable landing pages or an unrealistic budget. It also does not make every additional query profitable.
For a smaller South African account, control and learning matter more than being first. Start with a campaign that has trustworthy conversion history, a defined audience and enough data to evaluate. Keep a stable control and scale only when qualified business outcomes improve.
Sources
- Google Ads Help: How AI Max for Search campaigns works
- Google Ads Help: About AI Max experiments
- Google Ads Help: Reporting in AI Max for Search campaigns
- Search Engine Land: Ginny Marvin clarifies AI Max and AI Search Ads