Key takeaways
- Google’s AI-era Search ads add a Gemini-written explainer next to your creative, so your feed and landing page effectively become your ad copy.
- Business Agent for Leads swaps static forms for a chat agent, which lifts enquiry volume but risks counting low-value conversations as conversions.
- Native checkout for UCP merchants closes sales inside Google’s surface, weakening your own attribution and inflating already-generous platform ROAS.
- Access to these formats requires AI Max and Performance Max, meaning broader matching and less query visibility.
- Fix feeds, offline conversion tracking and query hygiene now, before these formats scale into your account.
AI-era Search ads are Google ad formats that use Gemini to interpret a searcher’s question, select relevant ads or products, and add AI-written context to sponsored placements. For UK advertisers, the immediate job is to clean product feeds, conversion tracking and account controls, because the model can only work with the data it can read.
Google’s newer Search and Shopping formats do not simply automate bidding. They change who shapes the message a buyer sees before clicking. In AI Mode, Shopping and lead formats, Google can generate explanations, answer questions and, in some commerce tests, move the transaction closer to Google’s own surface. That is useful when your data is accurate. It is expensive when your feed, website or conversion signals are vague.
Our view is direct: the accounts that win with AI-era Search ads will have clean commercial data, qualified conversion tracking and tighter human review. The accounts that lose will mistake extra automation for extra strategy. For more context on the wider shift, see our guide to Ads in AI Mode for UK PPC teams.
AI-era Search ads: what Google changed
Google’s Google Marketing Live Search ads announcement confirmed several changes that matter to advertisers. In AI Mode, Google described new ad experiences where sponsored results can appear inside AI-generated responses and recommendation-style results. These ads remain labelled as sponsored, but the surrounding context may be generated by Gemini rather than written directly by the advertiser.
Google also expanded AI-powered Shopping experiences, where the system can match products to a user’s expressed need and produce AI-generated reasoning about why a product is relevant. For lead generation, Google discussed Business Agent experiences that can answer customer questions using information from a company’s website. For commerce, Google described Direct Offers and checkout-related tests with selected partners.
The practical change is simple: in AI-era Search ads, your website, feed and structured data increasingly become the raw material for ad messaging. Headlines, descriptions and assets still matter, but the model’s interpretation of your business now matters as well.
Google has also positioned these formats alongside AI Max for Search, AI Max for Shopping and Performance Max. Its separate AI Max for Search guidance makes the direction clear: broader query coverage, more automated asset generation and more reliance on signals beyond traditional exact keyword targeting.
AI-era Search ads and ROAS: why the inputs matter
The first risk is not that AI writes poor copy. The first risk is that AI writes plausible copy from poor inputs. If your product title says one thing, your landing page says another and your Merchant Center attributes are incomplete, an AI-generated explainer can amplify the inconsistency at the point of decision.
In ecommerce accounts, we often see product feeds that were built years ago for basic Shopping visibility, then left alone. They may include thin titles, missing attributes, outdated promotions or product descriptions written for internal catalogues rather than customer search behaviour. Under classic Shopping, that was already a problem. Under AI-era Search ads, it becomes a messaging problem because the system may use those fields to explain your product to the buyer.
A product feed is no longer just campaign plumbing. In AI-led Shopping, the feed is part of the sales pitch. Google’s own Merchant Center product data specification shows how many fields can influence eligibility, matching and presentation, from titles and descriptions to availability, GTINs, pricing and shipping details.
Business Agent-style lead experiences create a different ROAS risk. A chat agent can lower friction and raise enquiry volume, but volume alone does not pay wages. If the agent answers the wrong questions, attracts low-intent conversations or records soft interactions as conversions, Smart Bidding can optimise towards activity rather than revenue. We see this pattern regularly in lead-generation accounts: more form fills, more chat starts, weaker sales pipeline. Our guide to small business lead generation services explains why qualified lead definitions matter more than headline enquiry counts.
AI-era Search ads and native checkout measurement
Checkout and offer experiences inside Google surfaces can reduce friction for merchants, but they also increase the importance of independent measurement. If more of the discovery, persuasion and transaction path sits inside Google’s environment, advertisers need a clear way to reconcile Google-reported value with their own revenue records.
Our Q1 2026 UK Google Ads spend analysis found that about 34% of UK Google Ads spend in the accounts reviewed went to Performance Max, with reported ROAS of 4.57 compared with Search at 3.15 on a like for like basis. That gap should not be read as proof that Performance Max is always more profitable. It can include brand demand, Shopping cannibalisation and platform-attributed conversion value that still needs commercial validation.
The more Google generates the journey and reports the result, the more advertisers need an independent source of truth. That is why we recommend reconciling platform ROAS against CRM, ecommerce platform or finance data before scaling new automated formats. Our cross platform measurement framework gives a practical structure for doing that.
AI-era Search ads in real accounts: PPC Geeks’ view
The issue we expect many advertisers to hit first is explainer mismatch. Imagine a client selling a mid-range espresso machine, positioned around ease of use and reliability. The feed emphasises pressure bars, boiler wattage and technical features. Gemini reads those details, then frames the product for a serious home barista rather than a busy family buyer. Clicks still arrive, but they come from the wrong audience. Conversion rate drops, Smart Bidding reads weaker demand and spend starts to move elsewhere.
Nothing in the interface is likely to say, ‘AI explainer drifted from commercial positioning’. It will show up as poorer click-through rate, weaker conversion rate or less stable bidding. That makes diagnosis harder, especially where broad match, Performance Max or AI Max already reduces query visibility.
In our audits, the highest-risk accounts usually share three traits: thin conversion history, broad automated targeting and weak feed or landing page hygiene. That combination gives automation more freedom while giving the advertiser fewer reliable signals to judge it.
Once a model writes context around your ad, your feed and landing page become ad copy. Most advertisers still treat them as back-office data, and that is where performance leaks start.Dan Trotter, PPC Director, PPC Geeks
This is exactly the type of issue we look for in a free Google Ads audit, especially where automation, feed quality or conversion tracking may be quietly shaping what AI-era Search ads say about a business.
AI-era Search ads preparation: what advertisers should fix now
Do not wait for these formats to appear fully in your account. Prepare the inputs now, because those inputs are what the model reads.
- Rewrite your feed for machine reading. Open Merchant Center, sort by impressions and audit your top 50 products. Improve titles, complete relevant attributes, remove unsupported claims and make sure product descriptions reflect how buyers search. If your catalogue is a major revenue driver, our guide to product feed management covers the areas to prioritise.
- Force feed and landing page agreement. Compare your ten highest-spend products side by side in Merchant Center and on the live site. Price, availability, delivery claims, returns information and positioning should match. Any contradiction is a future AI explanation problem.
- Separate lead quality from lead volume. Set up offline conversion tracking so a form fill, call or chat only trains bidding when it becomes a qualified opportunity or sale. Google’s offline conversion import documentation explains the data requirements, and our offline conversion tracking guide walks through the commercial setup.
- Check conversion import lag. In Google Ads, review conversion actions and upload timings. If qualified revenue appears several days after the click, Smart Bidding may be learning from incomplete data. Delayed signals are especially risky when new automated formats start spending quickly.
- Review AI Max and broad match query quality every week. Pull a 30-day search terms report where available, add negatives for irrelevant themes and check whether budget is drifting into low-intent research queries. Broader matching is useful only when waste is actively controlled.
- Reconcile platform ROAS with your own revenue. Compare Google-reported conversion value with Shopify, WooCommerce, CRM or finance data for the same period. If the gap widens as automated formats scale, fix measurement before adding budget.
If this list feels heavy for an in-house team, that is a sensible signal to bring in specialist support. Our Google Ads agency team runs these checks as standard across client accounts.
AI-era Search ads: what this means for UK campaigns
AI-era Search ads are not a threat to advertisers who prepare. They are a threat to sloppy inputs. Gemini-generated explanations, chat-led lead capture and checkout tests all reward accounts with accurate feeds, consistent landing pages and conversion tracking that measures commercial outcomes rather than motion.
The commercial reality is straightforward. More of the search journey is moving into Google’s automated and AI-mediated layers, and more of the reporting may depend on platform-defined conversion paths. If you cannot independently verify ROAS, you risk scaling spend on a number that looks precise but does not match profit.
For UK advertisers, the next move is not panic. It is maintenance with commercial discipline: clean the feed, align the landing page, qualify the conversion, review query quality and reconcile revenue. Do that before AI-era Search ads reach full scale, and the loss of manual control becomes manageable rather than costly.
We can help you stress-test your account against this. A free PPC audit is the fastest way to see where you stand: request your free Google Ads audit. For wider market context, the IAB UK Digital Adspend 2025 report shows how much budget is already flowing through digital advertising, with UK digital ad spend reaching £40.5bn in 2025.
Frequently asked questions
What are Google’s AI-era Search ads?
They are new ad formats built with Gemini, announced at Google Marketing Live 2026. They include Conversational Discovery ads and Highlighted Answers in AI Mode, AI-powered Shopping ads, Business Agent for Leads, and an expanded Direct Offers pilot. Most feature an AI-written explainer displayed next to your creative.
Why does the AI explainer matter for my ads?
Because Gemini synthesises your product from your feed and website, then writes context that sits inside your paid ad. If your feed is thin or contradicts your landing page, the explainer inherits those errors. You lose control of part of your ad copy and cannot A/B test it in the usual way.
How do AI-era Search ads affect ROAS reporting?
Native checkout completes purchases inside Google’s surface, so more revenue and value data flow through Google’s own reporting. Platform ROAS is already flattered by cannibalisation and unproven incrementality, so scaling these formats without reconciling against your own finance numbers risks trusting an inflated figure.
Do I need Performance Max to use these formats?
Yes. Google states you should build on AI Max for Search, AI Max for Shopping and Performance Max to access the new AI-powered formats. That means accepting broader matching and reduced query visibility as the price of entry.
What should UK advertisers do first?
Rewrite your product feed for machine reading, align feed data with landing pages, set up offline conversion tracking so chat-agent leads only count when qualified, check conversion import lag, and add negatives weekly on AI Max campaigns. Prepare the inputs before the formats reach your account.
