blog.google via Reddit

Google launches conversational shopping ads for AI Search

google search generative ai ai-advertising search agentic-commerce

Key insights

  • Google's new Shopping ads generate custom product explainers per query, replacing fixed creatives with real-time AI synthesis.
  • Business Agent for Leads lets users chat inside ad units, eliminating the external site handoff that has defined search ads.
  • Ask Advisor is a Gemini agent spanning Ads, Analytics, Merchant Center, and Google Marketing Platform as a unified optimization layer.

Why this matters

Google is shifting the fundamental unit of search advertising from a click to a conversation, which compresses the purchase funnel entirely inside its own surfaces and reduces advertiser control over the customer journey. The Agent Payments Protocol and Universal Commerce Protocol signal Google building durable infrastructure for agentic commerce, meaning third-party platforms that currently capture post-click value now face a structural bypass. For AI practitioners building on or competing with Google's stack, the Ask Advisor architecture shows how Gemini is being deployed not as a feature but as a cross-product orchestration layer that can ingest signals from multiple enterprise tools simultaneously.

Summary

Google used its May 20 Marketing Live event to roll out ad formats designed specifically for AI Search, moving beyond static display units into interactive, agent-driven commerce experiences. The centerpiece is a new class of AI-powered Shopping ads that generate custom product explainers on the fly, tailored to the specific query a user just typed. Alongside that, Business Agent for Leads lets users open a chat session inside an ad unit itself, getting instant answers without navigating away from the results page. On the advertiser side, Ask Advisor is a Gemini-powered optimization agent that spans Google Ads, Analytics, Merchant Center, and Google Marketing Platform simultaneously. Essentially: Google is rewiring its ad stack around two protocols, Agent Payments Protocol and Universal Commerce Protocol, that it's now extending into new verticals. - Shopping ads no longer serve a fixed creative; they synthesize a product explanation at query time based on what the user asked. - Business Agent for Leads keeps users inside the Google results page, reducing the click-to-site handoff that has defined search advertising since 2000. - Ask Advisor operates as a persistent optimization partner across the full Google marketing stack, not a one-off recommendation tool. The underlying bet is that keeping users inside Google's surfaces longer, via agents and in-ad conversations, is more valuable to advertisers than driving traffic to external sites.

Potential risks and opportunities

Risks

  • Retailers and D2C brands that rely on Google search traffic for first-party data acquisition could see click volume drop materially within 12 months if in-ad agents satisfy purchase intent before users reach brand sites
  • Advertisers using Business Agent for Leads hand conversational data to Google infrastructure, creating a dependency risk if Google uses those interaction signals to improve its own shopping surfaces at brand expense
  • Publishers and affiliate networks downstream of Google search face accelerated traffic erosion if agentic in-SERP commerce captures transactions that previously converted off-platform

Opportunities

  • Martech vendors with clean product feed infrastructure (Feedonomics, DataFeedWatch) gain leverage as the quality of real-time product explainers will be directly constrained by feed data quality
  • Google Shopping and Performance Max agencies with Gemini API access can build Ask Advisor integrations ahead of broader rollout, locking in advisory relationships before self-serve tooling commoditizes the workflow
  • Brands with strong structured data and schema markup are better positioned to surface accurate agent-generated explainers, giving SEO-forward organizations a short-term advantage over competitors with thin product data

What we don't know yet

  • Whether Business Agent for Leads stores or uses in-ad conversation data for ad targeting, and under what privacy policy disclosures
  • Which verticals are included in the first expansion of Agent Payments Protocol and Universal Commerce Protocol beyond the May 20 announcement
  • Whether Ask Advisor has access to raw Analytics event data or only aggregated metrics, which would determine its actual optimization ceiling