Google Ads AI is the set of automated features built into Google Ads: Smart Bidding sets bids, AI Max for Search and Performance Max choose queries and placements, Asset studio generates creative, and Ask Advisor answers questions and applies changes. Turn them on in that order, each one only after the one before it runs on clean conversion data.
This guide is part of the Google Ads hub and covers features inside Google Ads only. For AI across Meta, TikTok and ChatGPT, see the broader guide to AI ads. Product facts are as of October 2026 and come from Google's help center and announcements; Google's own performance claims are marked as such. The decision rules and the sequence below form what this guide calls the Google Ads AI rollout order.
Which Google Ads AI features exist as of October 2026?
As of October 2026, Google Ads AI covers six feature groups: Smart Bidding, AI Max for Search, Performance Max, Demand Gen, generative creative in Asset studio, and the Ask Advisor agent with the older conversational experience. Auto-applied recommendations sit on top of all six and can change account settings for you.
- Smart Bidding is a group of bid strategies that, in Google's words, "use Google AI to optimize for conversions or conversion value in every auction": Target CPA, Target ROAS, Maximize conversions and Maximize conversion value.
- AI Max for Search is a set of features inside a Search campaign, not a campaign type. Google's help page calls it "a continuous optimization layer for your Search campaigns."
- Performance Max is a goal-based campaign type that reaches all Google Ads inventory, including YouTube, Display, Search, Discover, Gmail and Maps, per Google.
- Demand Gen is a campaign type for image and video ads on YouTube (Shorts included), Discover, Gmail and the Google Display Network, per its help page.
- Asset studio is the Google Ads workspace for creating, editing and storing images, videos and text.
- Ask Advisor is the Gemini-built agent that analyzes the account and proposes or applies changes.
What does each Google Ads AI feature automate, and what do you keep?
Each Google Ads AI feature automates one decision and leaves another with you. Smart Bidding takes the bid, AI Max and Performance Max take query and placement choice, Asset studio takes first drafts of creative, and Ask Advisor takes analysis. You keep the conversion goal, values, budget, exclusions and the approval of anything that runs.
The table is the core of the Google Ads AI rollout order. The last two columns are decision rules, not Google guidance.
| Feature | What it automates | Control you keep | Turn on when | Keep off when |
|---|---|---|---|---|
| Smart Bidding | Bid per auction from signals such as device, location, time and query | Bid strategy, target, budget, conversion goals | Conversion tracking is verified and volume is steady | Primary conversions are mixed with soft actions |
| AI Max for Search | Query matching beyond keywords, ad text, landing page choice | Brand and URL exclusions, locations of interest, per ad group toggles | Smart Bidding performs stably and the budget is not limited | The campaign is limited by budget or the site has pages you cannot exclude |
| Performance Max | Channel, placement, audience and asset combination | Goals, assets, audience signals, search themes, brand exclusions, negative keywords | Feed or assets are complete and Search already converts | You cannot accept less placement control |
| Demand Gen | Visual ad delivery on YouTube, Discover and Gmail | Assets, audiences, bidding goal | Budget can reach at least ten times the target CPA per day, per Google | You need direct response from search intent only |
| Asset studio and generated assets | Headlines, descriptions, images and video edits | Approval of every asset before it runs | Brand rules are written down and someone reviews output | Claims are regulated or need legal sign-off |
| Ask Advisor | Analysis, diagnostics, suggested changes | Approval of each change | You want faster diagnosis in an English-language account, per Google | You work in a manager account, where it is not available |
| Auto-applied recommendations | Selected bid, keyword, targeting and ad changes | Which types are opted in, at account level | A type is low risk, such as removing redundant keywords | A type changes match types, targets or targeting reach |
When should you switch a campaign to Smart Bidding?
Switch a campaign to Smart Bidding once conversion tracking is verified and the conversions you mark as primary are the ones the business values. Smart Bidding uses auction-time signals, so it is only as good as the goal it chases. Every later Google Ads AI feature inherits that goal, which is why bidding comes first.
Google lists the signals the bidder reads in each auction: device, physical and intent location, time of day and weekday, remarketing lists, ad creative, language, browser and the actual search query, per its Smart Bidding page.
The same page sets the evaluation bar. Google recommends judging results over longer periods with at least 30 conversions, and at least 50 for Target ROAS.
Pick the strategy by how your conversions differ in value. If every lead is worth about the same, Target CPA or Maximize conversions fits. If order values or lead scores differ, value-based bidding fits, and the Target ROAS guide shows how to derive a target from your margin. Which click receives credit for a conversion depends on your attribution model, so settle that before the bidder learns from it.
Should you turn on AI Max, Performance Max or Demand Gen?
Turn on AI Max, Performance Max or Demand Gen only after Smart Bidding runs stably, because all three widen reach and lean on the same conversion signal. AI Max widens an existing Search campaign, Performance Max adds every Google channel in one campaign, and Demand Gen adds visual inventory where people are not yet searching.
AI Max for Search has three parts, per Google's help page: search term matching through broad match and keywordless technology, text customization, and final URL expansion. Controls include locations of interest, brand inclusions and exclusions, and URL inclusions and exclusions. Queries it adds show an "AI Max" match type in the search terms report. Google warns that "AI Max won't be effective if campaigns are limited by budget."
The performance claim is Google's own. At launch on May 6, 2025, Google said advertisers "will typically see 14% more conversions or conversion value at a similar CPA/ROAS," and 27% for campaigns still mostly on exact and phrase match, based on Google internal data from 2025.
In April 2026 Google announced AI Brief, a Gemini feature with messaging, matching and audience guidelines written in plain language, rolling out in English for AI Max for Search first, then Performance Max and AI Max for Shopping.
Performance Max accepts search themes and brand exclusions. To compare creative inside it, use Performance Max asset testing.
Should you let the Google Ads AI generator write your ads and images?
Let the Google Ads AI generator draft ads and images when someone reviews every asset before it runs and your brand rules are written down. Asset studio and the image generator produce first drafts quickly, but Google leaves editorial responsibility with the advertiser, and regulated claims still need a human check.
Asset studio generates product scenes, edits images by prompt (background changes, object removal, border extension), builds horizontal or vertical videos, adds voiceovers and writes headlines and descriptions. Generated projects are kept for 55 days.
Prompted image generation has conditions. Google's generated images page requires an account that has run ads for at least 30 days, has a history of policy compliance, does not advertise in a sensitive vertical and has not completed Election Advertising Verification. Prompts work in English, Dutch, French, German, Italian, Japanese, Portuguese and Spanish. Every generated image carries a SynthID watermark, which Google describes as "an imperceptible, digital watermark that's resistant to manipulations of the image."
Inside campaigns, generation also happens without a prompt: text customization in AI Max and Performance Max writes ad text for the query. AI Brief messaging guidelines, such as never mentioning prices, are the control for that text once they reach your account.
Can Google Ads AI agents and recommendations change your account?
Google Ads AI agents can change your account only with your approval: Ask Advisor proposes a change and applies it once you confirm. Auto-applied recommendations work differently. Once you opt in to a recommendation type at account level, Google applies eligible changes on a daily cycle without asking each time.
Google introduced the agent as Ads Advisor on November 12, 2025. Dan Taylor, Vice President of Global Ads, wrote: "With your review and approval, it can then apply these changes directly to your account." At Google Marketing Live on May 20, 2026, Google presented Ask Advisor, which "orchestrates a team of expert agents" across Google Ads, Google Analytics, Merchant Center and Google Marketing Platform.
The Ask Advisor help page calls it "an agentic conversational experience built within Gemini capabilities." It answers account questions, builds reports, troubleshoots performance drops and disapprovals, suggests keywords and assets, and makes individual changes such as pausing a campaign or updating a budget. It is in beta and not available in manager accounts. The older conversational experience still builds Search campaigns from a chat in English, French, Spanish or German.
Auto-applied recommendations cover select types in bidding, keywords and targeting, ads and measurement. A queue shows what will apply that day, and the History tab shows what did.
In what order should you roll out Google Ads AI?
Roll out Google Ads AI in four stages: data, reach, creative and delegation. Fix conversion goals and move to Smart Bidding first, widen reach with AI Max or Performance Max second, add generated creative third, and hand actions to Ask Advisor or auto-apply last. Each stage should prove itself in an experiment before the next starts.
- Audit conversion goals. Mark only business outcomes as primary and check that values match what each conversion is worth.
- Move to Smart Bidding. Pick Target CPA or value-based bidding and judge it on at least 30 conversions, or 50 for Target ROAS.
- Set exclusions before reach. Add brand exclusions, negative keywords and URL exclusions so wider matching stays inside the boundaries.
- Test AI Max or Performance Max on one campaign. Run it as an experiment with an end date and compare conversions on the same goal.
- Add generated assets with review. Use Asset studio drafts, approve each one, and test them against existing assets.
- Delegate last. Use Ask Advisor for diagnosis, approve each change yourself, and opt in only to low-risk auto-apply types.
- Check incrementality. Run a Google Ads Conversion Lift study when a stage claims a large gain.
Common mistakes with Google Ads AI
The common mistakes with Google Ads AI come from switching on reach and delegation before the data and controls are ready. Each mistake makes the system optimize toward the wrong goal or spend outside the boundaries you would have set by hand, and most are fixed by following the rollout order.
- Turning on AI Max before bidding is stable. Wider matching on a weak conversion goal buys more of the wrong traffic. Fix goals and bidding first.
- Running AI Max on a budget-limited campaign. Google says AI Max will not be effective there. Raise the budget or test on another campaign.
- Accepting Google's uplift figures as your forecast. The 14% and 27% AI Max figures in Google's May 2025 announcement come from Google internal data from 2025. Measure your own result against a control.
- Leaving auto-apply types unreviewed. Opted-in types keep applying daily. Review the History tab monthly and switch off types that change targets.
- Publishing generated assets unread. Asset studio drafts can state things your brand cannot promise. Approve every asset and keep regulated claims manual.
- Changing many AI settings at once. Change one setting per experiment, and use Google Ads Editor for bulk edits only after a test has a clear winner.
