Google Ads has launched a streamlined experiment format for Search campaigns called AI Max Experiment, allowing advertisers to test AI-based enhancements directly within an existing campaign. It divides traffic evenly—50% control vs. 50% test—without creating a campaign copy.
AI Max Experiment
▼ Why It Matters
Traditional experiment setups require cloning campaigns, causing delays and setup issues. The new AI Max approach simplifies testing and speeds up statistically significant results
_______________________________________
🚀 What It Enables
• Located under the “Choose a variable to test” section, you can now toggle AI Max for Search campaigns.
• It automatically activates:
o Search Term Matching (broadens query coverage)
o Asset Optimization (generates dynamic text assets)
• You can selectively disable these features at either campaign or ad group levels.
• Auto apply of successful test results is the default—with an option to override for manual rollout ________________________________________
🧠 Behind the Features
• Search Term Matching uses AI to interpret intent and match off keyword searches, helping reach broader yet relevant audiences.
• Asset Optimization builds headlines and descriptions using content from landing pages, blog text, and keywords—often outperforming static ad copy for better CTR and conversions
________________________________________
📊 How to Set It Up
1. Open the Experiments interface and click the plus (+) icon.
2. Choose “Campaign features and settings”, then select AI Max for Search campaigns.
3. Customize settings or use defaults for both main experiment levers.
4. Choose the 50/50 budget split and whether to auto apply results after completion.
5. Run the test and review performance via the Experiments dashboard
________________________________________
📈 Early Results and Use Cases
• L’Oréal doubled conversion rate while cutting cost-per-conversion by 31% using AI Max, tapping into new search queries related to their products
• MyConnect generated 16% more leads at a 13% lower CPA, attributing a 30% lift from previously untapped search terms
• Brands targeting multiple locations benefit from location intent targeting, reducing the need for separate campaign structures in local markets
________________________________________
⚠️ Limitations to Note
This setup can’t be used if your campaign:
• Targets the Display Network
• Uses shared budgets or Portfolio bidding
• Has legacy features like brand exclusions or text customization already enabled
• If needed, you can still opt for the original Custom experiment setup with separate campaigns
________________________________________
🧩 Best Practices for Testing AI Max
• Start small: Test in a single campaign before scaling.
• Exclusion rules: Set brand and URL filters upfront to avoid off-brand matches.
• Asset review: Periodically audit AI-generated headlines and descriptions.
• Track properly: Use Google’s {synthetic_keyword} parameter in your URL to segment AI Max specific traffic
________________________________________
✅ Summary Table
| Feature | Traditional Experiment | Built-In AI Max Experiment |
|---|---|---|
| Setup complexity | High (requires campaign duplication) | Low (within same campaign) |
| Budget split | Often manual setup | 50% control / 50% test by default |
| Features activated automatically | No | Yes (search matching + asset optimization) |
| Learning period | Longer | Shorter, faster insights |
| Sync & setup errors | More potential duplicate/configuration issues | Reduced risk within one campaign |
________________________________________
❓ Frequently Asked Questions
1. What is AI Max Experiment?
A built-in experiment in Google Ads for testing AI Max features within an existing Search campaign using a 50/50 budget split.
2. Which features are involved by default?
Search Term Matching (intent-based reach) and Asset Optimization (auto-generated ad creative).
3. Can I tweak which features to test?
Yes—you can disable Search Term Matching or Asset Optimization at the campaign or group level.
4. Will my campaign risk performance loss?
Only half the traffic is exposed to AI Max, so the control side keeps existing settings intact while testing proceeds safely.
5. What if I want manual result application?
Turn off the auto-apply feature when setting up the experiment.
6. Are there specific limitations?
Yes—campaigns using shared budgets, display targeting, portfolio bidding, or legacy brand/text features aren’t eligible.
________________________________________
💡 In Closing
Google’s new AI Max experiment setup cuts the overhead and complexity of running search campaign tests. By embedding testing within the live campaign and leveraging a 50/50 split, advertisers can gain statistically valid insights faster—with minimal setup risk. If you’re focused on efficiency and performance, AI Max is poised to be a powerful addition to your toolkit.
