Agentic AI is transforming shopping – here’s how performance marketers can win in the next 12 months

Search isn’t a list of links anymore. It’s becoming an agent that understands, reasons, and acts. From shortlisting products to completing purchases, Google’s rollout of Agentic AI marks the biggest transformation in retail performance since the launch of Smart Bidding – and retail will feel it first. Following our visit to Google’s Dublin HQ, we sat down with Catharina de Vlieg, Senior EMEA Product Lead in AI Solutions, to understand how Agentic AI will change the customer journey – and what it means for digital and performance leaders.  

Search is turning into a shopping assistant

AI Mode is changing the nature of discovery. It blends Gemini’s reasoning power with Google’s Shopping Graph – now over 50 billion product listings refreshed more than 2 billion times per hour – to turn simple queries into contextual, conversational shopping experiences. Instead of typing “best waterproof jacket” and scrolling through links, users can now ask “what should I wear for hiking in the rain in May?” and get a browsable, personalised panel of products that updates as preferences evolve. Discovery is no longer about keywords; it’s about context, conversation, and intent.

And that’s just the start

Google’s Agentic Checkout will allow shoppers to set their ideal price and hand the purchase off to Google Pay – a “buy for me” feature that completes transactions automatically once the target price is hit. Virtual try-on (VTO) lets users see how clothes drape on their own photo. And visual search – through Lens and Circle to Search – now drives 25 billion+ searches every month, many with clear commercial intent. For performance marketers, these changes mean one thing: control is shifting. The AI layer now decides what to show, when to show it, and how much of that journey ever reaches your website.  

Challenges for performance marketing leaders

For performance marketing leaders, Agentic AI isn’t just another feature update – it’s a shift in how performance is defined and measured.

• Attribution will fragment: Agents will complete purchases inside Google’s ecosystem, meaning fewer trackable clicks and fewer sessions that look like “conversions.”
• Feed quality becomes your ranking engine: Your product data, structured schema, and on-page content now dictate whether you appear in AI Mode panels – not your keyword coverage or bid strategy.
• Creative becomes measurable: AI will only surface products it can “reason about.” If your imagery, copy, and schema don’t answer the question behind the query, you won’t be chosen.
• Performance becomes a systems challenge, not a channel challenge: The job isn’t to optimise ads in isolation anymore – it’s to optimise the data, content, and creative signals that AI can interpret as relevance and trust.

 

How to get ahead now

Start with your feed

Your Merchant Center feed and PDPs now work like a combined ranking engine for AI Mode. Make sure product titles, attributes, and structured data are clean, complete, and consistent. Fix contradictions between feeds and pages. Enrich listings with reviews and real user-generated content that the AI can cite as trustworthy.

Prepare your site for agents, not just users

Integrate Google Pay, enable guest checkout, and ensure your cart and product endpoints are accessible to Google’s crawlers. Avoid blocking legitimate bots – these are now how AI models “see” your storefront.

Make creative your competitive edge

AI will favour the listings that communicate clearly – through high-resolution images, descriptive alt text, and short, benefit-led copy. Think “answer density” over adjectives.

Redefine your campaign architecture

Performance media still matters – but how it performs is changing. Ad placements are already appearing within AI Overviews and will soon expand to AI Mode. Eligibility will depend on feed completeness and contextual relevance, not just keyword match. This means embracing broad match and Smart Bidding to capture the long-tail, AI-generated queries no one can predict, while using Performance Max / AI Max to reach consumers in the new discovery layers before they ever click a link. Demand Gen can then bridge the gap, reaching upper-funnel audiences in YouTube, Discover, and Gmail before AI Mode narrows their options. In short: campaign architecture needs to mirror the new customer journey – from inspiration to action without interruption.

Rethink measurement before it breaks

Agentic AI will compress and blur the funnel. The classic clickstream is dissolving. Strengthen your first-party event tracking (view_item_list → add_to_cart → purchase) and modelled conversions to handle fewer direct signals. Treat micro-actions like price tracking, “buy for me” triggers, and agent-assisted checkouts as new conversion events. This is also the time to build or refresh your MMM or calibrated incrementality models – so you can attribute value when conversions happen inside Google’s ecosystem, not just on your site.  

The next 12 months: performance redefined

Agentic AI won’t replace performance marketers – but it will expose inefficiency. The marketers who win won’t be the ones with the biggest budgets or the longest keyword lists. They’ll be the ones who:

• See AI not as automation, but as a new discovery surface.
• Invest in data accuracy and creative quality as growth levers.
• Align feeds, content, and campaigns as one system – not separate teams.

  The brands that win will be the ones built to show up, stand out, and stay relevant when discovery happens everywhere.   Want to see how these insights could benefit your brand?  Connect with our team.