If you run Google Ads, September 2026 is not a month to leave the account on autopilot. Google has begun automatically migrating campaign-level broad match and automatically created assets into AI Max for Search campaigns, and as of early September it no longer allows new legacy structures to be created. Dynamic Search Ads follow in early 2027. This is the largest structural change to Search campaigns since Performance Max launched, and the accounts that come out ahead are the ones that audit before the migration lands rather than after. Here is what is actually changing, what you keep control of, and the specific work we are doing in client accounts this month.
Key Takeaways
- Google Ads AI Max replaces traditional keyword match types with intent-based targeting across search campaigns
- AI Max campaigns require a minimum of 30 days learning period - expect CPA spikes in the first 2-4 weeks
- Creative assets (headlines, descriptions, images) become the primary lever for performance in AI Max - invest in variety
- Negative keyword lists and brand exclusions remain critical controls - set these before enabling AI Max on any campaign
- Early adopters of AI Max in India report 15-25% improvement in conversion volume at similar CPA after the learning period
The Migration Timeline You Need on the Calendar
Google has published a staged timeline rather than a single switch-over date. Campaigns are migrated in place using equivalent AI Max settings, and existing brand inclusions and exclusions carry over automatically - but "carried over automatically" is not the same as "configured correctly for your business".
- Through September 2026 - automatic migration of campaign-level broad match and automatically created assets (ACA) to AI Max
- Early September 2026 - creation of new legacy structures is switched off; anything you were planning to build the old way has to be built now or built differently
- September 2026 onward - in-account notices begin encouraging voluntary Dynamic Search Ads upgrades
- Mid-January 2027 - reminder notifications about the DSA migration
- February 2027 - automatic DSA migration; the ability to create new DSA ad groups is permanently removed
- Around September 2027 - older API versions supporting legacy features reach sunset, which matters if you run scripts or third-party integrations
What AI Max Actually Changes in the Account
AI Max moves Search campaigns from a keyword-list model toward an intent-matching model. Google expands the queries a campaign can serve against using your landing pages, assets and existing keywords as signal, then generates and assembles ad copy to match. The practical effect is that the boundary of your campaign is no longer defined solely by the keywords you typed in. For a business with wide margins and broad appeal, that is usually upside. For a business selling one specific high-ticket service in one region, it is a budget risk until it is constrained properly.
The Controls You Still Have - And Should Be Using
The single most common mistake we see is treating AI Max as all-or-nothing. It is not. The controls that protect spend are still there, and they matter more now than they did under manual keyword targeting, because the system is generating reach you did not explicitly ask for.
- Negative keywords - still the primary lever. Lists should be rebuilt before migration, not after you have paid for the search terms that taught you.
- Brand inclusions and exclusions - these carry over, but audit them; an exclusion list built for broad match may not cover the queries AI Max reaches.
- Location targeting - tighten to your actual service area. Presence-based targeting rather than presence-or-interest is usually the right setting for local service businesses.
- Final URL expansion - decide deliberately whether Google may send traffic to pages you did not nominate. For most lead-gen accounts, restrict it.
- Asset-level controls - pinning and asset exclusions still govern what the system can assemble.
- Audience signals - feed it your converter lists and customer match data so the expansion starts from good examples rather than guesses.
Why Conversion Tracking Quality Decides Whether AI Max Works
Every automated bidding and matching system is a function of the conversion data you feed it. Under manual keywords, bad tracking produced mediocre results. Under AI Max, bad tracking produces confidently wrong results at scale, because the system is expanding reach in the direction your conversion data points. If you are counting every form fill as a conversion - including the spam, the job applicants and the vendors - AI Max will find you more of exactly that. Before migration, we make sure conversions represent qualified business outcomes: offline conversion import from the CRM where the sales cycle is long, value-based bidding where project sizes differ materially, and conversion actions that exclude the low-quality paths.
What We Are Doing in Client Accounts This Month
The pre-migration audit is not complicated, but it has to happen before the migration rather than after. This is the sequence we run.
- Export 12 months of search terms and build an expanded negative list from everything that produced spend without qualified enquiries
- Verify every conversion action actually represents a business outcome, and remove or downgrade the ones that do not
- Push offline conversion data back to Google Ads for any account with a sales cycle longer than a week
- Tighten geography to real service areas, switching to presence-based targeting
- Decide final URL expansion deliberately per campaign and document the decision
- Set a clean pre-migration performance baseline - CPL, qualified lead rate, cost per qualified lead - so post-migration change is measurable rather than anecdotal
- Diarise a structured review at 14 and 30 days post-migration, because the learning period will move numbers before it settles them
If You Run Scripts, an API Integration, or a Third-Party Tool
The API sunset is the part that tends to get missed until something breaks silently. Older API versions supporting legacy features are scheduled to sunset around September 2027, which sounds distant until you account for how many reporting dashboards, bid scripts and CRM integrations were built once and never revisited. Audit now: list every script and integration touching the account, identify which API version each depends on, and schedule the updates. A reporting dashboard that quietly stops updating is worse than one that visibly breaks.
Frequently Asked Questions
What is Google Ads AI Max and how does it differ from Performance Max?
AI Max is Google's next-generation campaign type launched in mid-2026, combining the broad reach of Performance Max with significantly improved controls for advertisers. Unlike Performance Max, AI Max lets you see which search terms triggered your ads, exclude specific placements more precisely, and retain more brand safety controls. It uses Google's Gemini AI models for creative generation and bidding, with better attribution than its predecessor.
Should I migrate my existing campaigns to AI Max immediately?
Not necessarily immediately. Google is running AI Max alongside existing campaign types through late 2026, with a planned full transition by 2027. The recommended approach is to run parallel tests: keep your best-performing existing campaigns while testing AI Max with a portion of budget (30-40%). Wait 4-6 weeks of data before making full migration decisions. Some campaign types - particularly branded search and high-intent exact match - may still perform better in traditional Search campaigns.
What creative assets are needed for AI Max campaigns?
AI Max requires a broader asset library than standard campaigns: minimum 5 headlines and 5 descriptions (15 recommended), at least 3 landscape images (1200x628px), 1 square image (1200x1200px), and 1 logo. For best results, include at least one 6-second bumper video and one 15-30 second video. The AI system generates ad variations from these assets, so more diverse input assets generally produce better output combinations.
How does AI Max bidding work and can I control it?
AI Max uses Google's latest Smart Bidding, primarily Target CPA or Target ROAS. You set the target, and the AI optimises bids across channels, times, and audiences to hit it. Controls available to advertisers include: URL expansion on/off, brand exclusions, topic exclusions, location targeting precision, and search term category reporting. You cannot manually set bids on individual keywords within AI Max - if granular keyword-level control is essential, keep a traditional Search campaign running alongside it.
Key Takeaway
AI Max is not optional and the timeline is not negotiable, so the only variable you control is how prepared the account is when the migration lands. The accounts that do well will be the ones with clean conversion data, disciplined negatives, honest geography and a documented performance baseline. The accounts that struggle will be the ones that let the migration happen to them and then try to diagnose a drop with no "before" to compare against. If you want a pre-migration audit on your account, we will run one and send you the findings - whether or not you work with us afterwards.
About the Author
Shweta Rohilla
Founder, AddMads
Shweta founded AddMads and leads strategy and growth. She writes about performance marketing, SEO and AI-first brand building for Indian and global markets.
