Google Ads AI Mode Tests Restrictive Match Types & 2026 Updates

The key change is that Google Ads has begun a small experiment allowing exact and phrase match keywords to serve text ads in AI Mode within pure Search campaigns. Google's Ads Liaison, Ginny Marvin, confirmed the test, stating that restrictive match types will only trigger when there is "explicit and direct user intent." This marks a departure from the previous requirement that AI-powered placements required AI Max, Performance Max, or broad match with smart bidding. Anthony Higman first spotted the change, and it has generated significant discussion among PPC professionals.

Google Ads Tests Restrictive Match Types in AI Mode

The experiment represents a middle ground: advertisers gain more control over which queries trigger their AI-driven ads, but only when user intent is clear. This contrasts with AI Max and PMax, which are designed for complex conversational intent and have access to newer ad formats like generative video and image assets. The table below summarizes the key differences:

Feature	AI Mode (Restrictive Match Test)	AI Max	Performance Max
Match Types Supported	Exact, Phrase (limited)	Broad match primarily	Broad match
Intent Handling	Explicit, direct user intent	Complex conversational	All intent types
Ad Formats	Text ads	Text + generative assets	All formats including video
Control Level	High (keyword control)	Low (automated)	Medium (some signals)

As reported by Marketers Index, the PPC industry sees this as a "significant shift" because it offers more precise targeting within AI-driven campaigns while retaining the benefits of automation. Advertisers running pure Search campaigns should test this feature proactively to understand how it affects impression share and cost.

The test is limited to text ads and only activates when Google's algorithm detects clear user intent. For example, a search for "buy blue running shoes size 10" is more likely to trigger an exact-match ad than a vague query like "running shoes." This nuance is critical for advertisers who fear losing control to broad match.

Target-Based Bidding: What Changed in August 2026

Google announced a change to target-based bidding for campaigns marked as "Limited by budget," effective August 17, 2026. According to Technotification, campaigns will now optimize more consistently toward the set target (Target CPA or Target ROAS) rather than over-performing beyond the goal when budget is constrained. This means if an account is budget-limited, Google Ads will no longer aim for a lower CPA than the target; it will likely spend the full budget while hitting the target exactly, potentially increasing costs per conversion if the target is too aggressive.

Google has provided a Bid Target Adjustment Tool to help advertisers estimate the impact and adjust targets accordingly. For example, if your campaign had a Target CPA of $50 but was achieving $40 due to budget constraints, after August 17 it will drift toward $50, raising your cost per acquisition. Advertisers running Target CPA or Target ROAS campaigns with limited budgets should review their targets and ensure they are realistic.

This change applies to all campaign types that use target-based bidding, including Search, Display, and Shopping. The impact can be sudden, so it's wise to check budget settings and targets immediately.

AI Max for Search: Default for High-Spend Accounts

Since late July 2026, Google has been making AI Max for Search Campaigns the default recommendation for accounts spending over $5,000 per month. This fully automated, generative-AI-driven ad format promises lower CPAs and higher conversion volumes. However, independent agency data suggests a "messier" reality for some brands, particularly those with niche products or tight creative control. Online Store News reports that the rollout has been polarizing, with some DTC brands seeing reduced efficiency.

AI Max leverages generative AI to create ad copy, images, and even video assets on the fly. For brands that value consistent messaging, this can be problematic. Advertisers should monitor performance closely and consider whether to opt-out if the default recommendation does not align with their goals. The forced push toward AI Max reflects Google's strategy to automate more of the ad management process, but it also removes some advertiser control over creative and targeting.

Restricted Ad Serving: A Growing Risk for Impressions

Google's "Restricted Ad Serving" policy, introduced in 2023 for Search and YouTube and expanded in 2026 to other surfaces, can limit an account's eligibility to participate in auctions, drastically reducing impressions without an account disapproval. XBorder Insights details a case study where a $3 million account lost 56% of its impressions overnight due to policy triggers. The policy is designed to protect user experience by limiting serving for accounts that have certain policy issues, but it can impact even compliant accounts if Google's algorithms flag them.

Advertisers experiencing sudden drops in impressions should review their account health, policy compliance, and the Limited Ad Serving documentation to understand potential causes and remedies. Common triggers include sensitive content flags, insufficient ad reviews, or changes in business information. Unlike an account suspension, restricted serving does not come with a clear notification, making it easy to overlook.

Other Notable Updates: Collapsing Sitelinks and Dismissed Recommendations

Google Ads is also testing collapsing sitelinks that users can expand, as reported by Search Engine Roundtable. This change affects how sitelink extensions appear in search results. Additionally, Google is pushing dismissed recommendations back into view, reminding advertisers of suggestions they previously declined. Both updates are part of Google's effort to streamline the ad interface and encourage adoption of its best practices. Collapsing sitelinks may reduce clutter but could also lower click-through rates for sitelinks. Meanwhile, pushing dismissed recommendations can be annoying but serves as a nudge to implement changes Google deems beneficial.

Practical Steps for Advertisers

The updates outlined above require proactive adjustments:

• Test Restrictive Match in AI Mode: If you run pure Search campaigns, enable this test to see if you can achieve better control while tapping into AI Mode placements.
• Review Target CPA/ROAS: With the bidding change, ensure your targets are realistic and account for the new optimization behavior. Use Google's Bid Target Adjustment Tool.
• Evaluate AI Max Impact: If your account is spending over $5k/month, monitor whether the forced AI Max recommendation improves or worsens performance. Consider reverting to standard campaigns if creative control is essential.
• Check for Restricted Serving: Regularly audit impression patterns and account policy status to catch restricted serving early.
• Leverage AI Prompts: Practical Ecommerce suggests using specific AI prompts to generate ad copy and strategies within Google Ads, which can help maintain efficiency. Their guide on 3 AI prompts for Google Ads offers actionable templates.

Additionally, some advertisers are exploring managing PPC outside Google Ads altogether. Search Engine Journal discusses what happens when you stop using Google Ads to manage your campaigns, highlighting third-party tools and automated scripts as alternatives. That article provides a useful counterpoint to the all-in-Google approach.

Conclusion

September 2026 is a pivotal month for Google Ads. The restrictive match type experiment in AI Mode could reshape how advertisers approach AI-powered search, while the bidding change and AI Max defaults require immediate attention. The restricted ad serving policy is a silent impression killer that demands vigilance. By adapting to these changes, advertisers can maintain efficiency and control in an increasingly automated ecosystem.