Skip to content
IFAI

AI tools

Google brings generative AI to Performance Max

Published Reach 36

Google just rolled out generative AI features in beta to all Performance Max advertisers in the U.S. Marketers can now automatically generate unique headlines, descriptions, and images with simple prompts, alongside in-app background editing. All AI-generated visuals will also carry invisible SynthID watermarks for safety and provenance.

Ask any media buyer about their biggest headache, and they won’t say budget management. They’ll say creative fatigue. Feeding hungry ad algorithms requires endless headlines, descriptions, and images tailored to every possible screen size.

Google wants its algorithms to solve that problem on their own. The company rolled out generative AI asset creation tools in Performance Max (PMax) as a beta, available to all US advertisers by the end of the year.

PMax launched back in 2021 to let marketers run a single campaign across Search, YouTube, Display, Discover, Gmail, and Maps. The catch has always been asset variety. If you don't supply enough high-performing creatives, your cross-channel reach falls flat.

Ad creatives on autopilot

With the new update, advertisers can prompt Google’s AI to churn out brand-new headlines, descriptions, and lifestyle images within seconds. Even better, PMax incorporates campaign performance data into its recommendations, actively nudging you toward assets likely to convert.

To address copyright and duplicate concerns, Google stated that its AI will never generate two identical images across different businesses or clients. Every output is unique.

Beyond generating visuals from scratch, an AI image editor inside Google Ads lets marketers swap backgrounds and tweak hero assets on the fly. An agency can turn a standard summer product photo into a snowy holiday scene without booking a new photoshoot. Image editing will roll out globally to all accounts by early 2024, alongside shareable preview links for client approvals.

On the safety side, Google is tagging every generated image with its invisible SynthID watermark and open standard metadata. Advertisers still have to review and approve assets before launch, and normal ad policy checks remain intact.

Why it matters

We think this is one of Google's smartest moves to lock ad dollars inside its ecosystem, especially from resource-strapped small businesses.

Historically, running omnichannel campaigns across half a dozen Google formats required dedicated design teams. Small businesses simply couldn't produce enough variations to make PMax work effectively. By letting marketers build full creative suites from a few text prompts, Google is eliminating that barrier to entry.

However, there is an obvious trap. When thousands of competing brands rely on the same foundational model to design their holiday promos, visual homogeneity is inevitable. Unique pixels do not guarantee fresh ideas.

For agencies and marketers, churning out minor image variations will no longer justify retainers. The real value is shifting upstream, toward prompt craftsmanship, strategic positioning, and knowing which generated concepts are actually worth putting budget behind.

Sources (1)

  1. [1]
    blog.google(opens in a new tab)Originalblog.google

Topic trackerGoogle's generative AI integration in advertising

Every story on this topic, newest first.

  1. Google Ads just got a Gemini-powered chatbot to build your campaigns
  2. Google brings generative AI to Performance MaxThis story
  3. Google gives merchants free AI product photoshoots

Best of AI, daily