Act Now on Google’s AI Search Shake-Up: Don’t go into peak on the Back Foot

By Josh Marinaro and Jon Breeds

Inside the Google ecosystem, an algorithmic shift is rewriting how consumers research, browse, and buy from e-commerce brands.

The critical backend mechanics of making your product data machine-readable still stand, and we’ve published a detailed guide to getting that right. But structural hygiene alone is no longer enough to protect your pipeline. Consumer behaviour has changed permanently.

To keep scaling brands ahead of the curve, Launch’s paid media experts, Jon Breeds and Josh Marinaro, sat down to unpack what these changes mean for your growth strategy.

From Keywords to Conversations: The 8-Word Search Avalanche

For years, digital marketing leaders have built acquisition strategies around predicting distinct, short-tail search terms. But the way real people look for products has evolved beyond basic exact-match indexing.

Lives are busier than ever, and consumers are increasingly letting AI handle the heavy lifting of research and decision making when making purchases. Think about it from a consumer standpoint: if you are targeted by ads telling you to buy supplements or mushroom gummies, you don’t just blindly click buy. You head to Google and ask the AI to compare the options, benefits, and price points for you in real time.

The data from recent closed-door Google events backs up this behavioural shift:

• Long-tail search queries containing eight or more words are up by a massive 800% year-on-year.
• Buyers are heavily relying on conversational, question-led prompts inside AI overviews, Gemini, and ChatGPT to guide their purchases .
  

This behavioural shift changes the entire creative template; your ads can no longer just tell someone to “buy this product”—they have to actively answer the consumer’s specific problem.

moving from keywords to context

This means exact-match keywords are becoming borderline impossible to predict.

Google highlighted a scenario involving an e-commerce brand selling espresso machines.

A consumer might spend weeks conducting high-intent brand research. However, their very final search (the moment they are holding a credit card and ready to buy) might be something seemingly unrelated, like “perfect housewarming gift for my neighbour” .

Historically, your account would have ignored that phrase or marked it as a negative keyword. Today, your brand needs an advertising setup built to understand that behavioural context and intent rather than just explicit text strings.

Taming the Machine with Brand Safety Guardrails

Allowing automated frameworks to dynamically expand your query mapping sounds incredibly efficient, but for premium or luxury e-commerce brands, it naturally sets off alarm bells.

No marketing director wants to risk an automated system generating ad variations that look completely off-brand, bizarre, or falsely position a luxury item under ‘cheapest’ messaging .

Fortunately, recent safety updates within Google’s AI Max performance tool have finally introduced strict commercial guardrails to protect your positioning:

• Custom Exclusion Prompts: Marketing teams can now feed up to 40 distinct guidelines directly into the AI engine.
• Strict Parameter Controls: You can explicitly command the system on what not to say—forbidding specific brand name misuses, pricing definitions, or generic catchphrases.
• Asset Previews: You can actively review mockups and served asset combinations built by the algorithm before they are exposed to live audiences.
  

These guidelines provide a vital safety layer that didn’t exist a few months ago, making automated expansion structures far safer to adopt.

YouTube is Dominating the Smart TV Storefront

While search intent becomes more conversational, video platforms are moving to eliminate purchase friction entirely.

YouTube is currently the largest streaming platform across both the UK and the US, fundamentally dominating viewer watch time over mainstream subscription networks like Netflix, Amazon Prime, and Disney.

Because it is a default software setting built directly into modern smart TV interfaces and physical remote controls, consumer accessibility has never been higher.

At Google Marketing Live we learned how video ads are transforming into transaction portals:

buy it now coming to streaming assets

An upcoming feature migrating to the UK and European markets introduces a native ‘Buy It Now’ layout directly beneath mobile and streaming video ad assets. When a user interacts with a compelling, high-quality video ad, the exact product thumbnails appear instantly underneath.

Clicking the link fast-tracks the user past site navigation pages and drops them directly onto the retailer’s absolute checkout page. This single-click execution style mirrors historical conversion optimisations that generated billions in incremental revenue for platforms like Amazon.

Smart Video Placements over Blind Audiences

Video advertising is often met with commercial anxiety because it can feel easy to dilute a budget across vast networks without hitting tangible returns. To maximise efficiency, brands should look at pairing contextual channel placements alongside standard audience lists.

For instance, if your brand sells premium kitchen knives or chopping boards, targeting specific, highly relevant video channels (like a cookery show or a Gordon Ramsay clip) ensures your message hits the right place at the absolute right time. Meeting your audience directly within their hyper-focused niches guarantees a much stronger commercial yield.

Creator partnership libraries

Furthermore, Google is launching creator partnership libraries in beta. This allows brands to search analytics data, identify content creators who are already naturally discussing their products, and back them with partnership funding, bypassing high internal video production costs entirely.

De-Risking the search Transition

The absolute worst mistake an e-commerce brand can make right now is treating these platform updates like a future problem. The algorithmic shift isn’t going backwards, and waiting means forcing your team to play catch-up on the back foot.

To embrace these changes safely without cannibalising your historical account performance, follow these two core execution rules:

1.Run a 50/50 Controlled Experiment

Never just blindly flip the switch on new automation settings across your entire catalogue. Instead, run a clean campaign experiment.

Keep your standard, proven campaign running as normal on 50% of your budget.

Allocate the remaining 50% to a duplicate test campaign with AI Max conversational expansion parameters enabled. This allows you to contrast performance behaviour and make data-backed growth decisions without taking a blind risk.

2.Invest Heavily in Asset Quality

Automation is reliant on the context you feed it. Ensure you are putting effort into compelling, relevant video creative that resonates emotionally with your market, rather than just relying on generic, dry product displays.

Are you set up for an AI powered search landscape?

If you want to ensure your performance marketing structures are genuinely optimised to scale alongside these behavioural intent changes before budgets face board-level scrutiny, get in touch with our paid media team today for a bespoke account audit.

About the authors

Josh Marinaro

Josh has 8 years of experience in performance marketing, specializing in PPC and CRO strategies. Before joining Launch, he worked as a PPC Manager at a full-service agency. At Launch, Josh focuses on combining his work across paid search and experimentation to get award-winning results for his ecommerce and travel clients.

Connect with Josh on LinkedIn.

jon breeds

Jon has 5 years’ digital marketing experience with particular expertise in Youtube ad placements. He’s worked with clients including Forthglade and Icelandair.

Connect with Jon on LinkedIn.