Just a few days ago, Google held its Google Marketing Live 2026 event, and if you manage campaigns or run a business that relies on online advertising, you\u2019d be well advised to take a moment to look closely at what was announced. This isn\u2019t just another update. It\u2019s a complete overhaul of the Google Ads playbook centred around Gemini, its AI model. The company has summed it up in a phrase it repeats on its own blog: they are rebuilding the marketing stack to make it \u201cAI-native\u201d. And that, in plain English, means that many of the things we\u2019ve been doing in Ads for years are about to change fundamentally.<\/p>\n\n\n\n
At Vandelay, we\u2019ve been analysing the official announcements and the initial trials already underway in the US for several weeks. This article summarises what we know for certain, what\u2019s still in beta, and what you should start preparing for now, even though some features may take months to roll out in Spain.<\/p>\n\n\n\n
Google Marketing Live is the annual event where Google shows advertisers, agencies and developers the direction its advertising platform is taking. Historically, it has been the venue for the launch of Performance Max, Demand Gen campaigns and improvements to Smart Bidding. This year\u2019s edition, however, has taken on a different focus: as reported by Search Engine Land, the event has revolved entirely around \u201cAI-powered agencies\u201d, conversational search, automated creative production and AI-assisted buying.<\/p>\n\n\n\n
To put it another way: until now, adverts coexisted with AI. From this edition onwards, adverts exist within AI. And that has very specific implications for how we bid, measure and, above all, create adverts.<\/p>\n\n\n\n
The announcement that has sparked the most discussion within our team is Ask Advisor. According to Google\u2019s documentation, it is a unified agent built using Gemini that acts as a cross-functional layer across Google Ads, Google Analytics, Merchant Center and the Google Marketing Platform. Until now, each of these tools existed in its own tab, with its own metrics, permissions and interfaces. Ask Advisor aims to bring them together into a continuous conversation.<\/p>\n\n\n\n
What does that mean in practice? It means that an advertiser will be able to ask, in natural language, questions such as \u201cWhich campaign has the most scope for scaling up next quarter?\u201d, and the agent will cross-reference data from the four platforms to provide an answer. Google describes it as an \u201calways-on strategic partner\u201d that helps optimise, diagnose and build campaigns without having to switch between interfaces. That\u2019s the promise, of course; it remains to be seen to what extent the agent understands complex account contexts. But the direction is clear: fewer clicks on the interface, more decisions made through conversation.<\/p>\n\n\n\n
Another format that has generated a great deal of excitement is Conversational Discovery Ads. Here\u2019s how it works: when someone makes a query in AI Mode (Google\u2019s conversational mode), Gemini instantly generates ad creatives tailored to the exact intent of that search and also writes a separate \u201cexplainer\u201d that summarises the context of the product or service.<\/p>\n\n\n\n
The example Google gives on its official blog is telling: someone searches for \u201chow to make my home smell like a spa\u201d and the system serves up personalised adverts featuring relevant products and an explanation of why they match that search intent. For an advertiser, this changes two things. Firstly, the creative stops being a fixed asset uploaded to the manager; it becomes a set of components (headlines, descriptions, images, benefits) that Gemini will recombine based on the query. Secondly, traditional copywriting loses ground to the \u201cfeed\u201d of information we provide to the system. The more structured information there is, the better the model will be able to explain the results.<\/p>\n\n\n\n
Search Engine Land confirms that these formats, along with Highlighted Answers, are already being tested in the US on both mobile and desktop. There is no fixed roll-out date for Spain, but everything suggests it will arrive sometime next year.<\/p>\n\n\n\n
If you run a service-based business, this is probably the change that will affect you the most. Google has announced Business Agent for Leads, a lead generation format in which the traditional static form is replaced by a conversation. A chat window appears within the advert itself, featuring a Gemini agent trained on the information from the advertiser\u2019s website. The user asks questions, the agent replies, and the conversation eventually turns into a qualified lead.<\/p>\n\n\n\n
Google has confirmed that the format is in open beta for advertisers in the US. The official example they give is that of a student researching universities: instead of filling in a form, they click on \u201cChat\u201d and get immediate answers about courses, scholarships and deadlines. What\u2019s interesting is what this implies for the rest of the funnel: if the agent filters out queries before generating the lead, the forms that reach the CRM will be much more qualified. But it also means you need to have a well-structured website and consistent messaging, because the model will pick up on any inconsistencies.<\/p>\n\n\n\n
The section on e-commerce updates is also worth noting. AI-powered Shopping Ads use Gemini to generate personalised explanations of why a product matches the user\u2019s search. Google\u2019s example: someone searches for an espresso machine and the system selects the most relevant models and writes a \u201ccustom explainer\u201d highlighting why that product might be the right one. It is not a generic sales pitch; it is text generated specifically for that particular search intent.<\/p>\n\n\n\n
Added to this is the \u2018Highlighted Answers\u2019 format, which highlights direct answers within AI Mode. For online shops, this reinforces the importance of having detailed product listings with all attributes properly filled in on Merchant Center, as this is the data Gemini uses to generate the text that users will see.<\/p>\n\n\n\n
This is where we need to get everything sorted. Most of these features will be rolled out in the US first and then in Europe; some will remain in beta for months. But there are decisions that are best made now, as they\u2019ll set the stage for what\u2019s to come.<\/p>\n\n\n\n
The first step is to review the quality of the feed and the assets. The richer and better structured the information on products, services and value propositions is, the better the generative creatives will perform. This includes attributes in Merchant Center, long descriptions in Performance Max assets and detailed content on the website itself.<\/p>\n\n\n\n
The second is to organise your website with the idea that an agent will be reading it. If Business Agent for Leads trains its chatbot using your site, anything that\u2019s poorly explained or out of date will result in a weak response right in the middle of the advert. This ties in with a concept we discussed a few months ago on the blog: llms.txt files and information prepared for LLMs aren\u2019t just about organic SEO; they now affect your ads too.<\/p>\n\n\n\n
The third step is to rethink how we measure things. If Ask Advisor cross-references data from Ads, Analytics, Merchant Center and GMP, it\u2019s important that these systems are properly connected and that the events are clean. Any gaps in GA4, or any duplicate conversions, will skew the agent\u2019s recommendations.<\/p>\n\n\n\n
There is one aspect of the official narrative that warrants further clarification. Google has been pushing for less control and more automation for years: Performance Max has already reduced control over ad placements; Smart Bidding has done the same with bids; and AI Mode reorganises the SERP. These new formats, with Gemini writing the copy and the agent deciding what is shown, reinforce this trend. For the advertiser, this represents a real trade-off: greater operational efficiency in exchange for less visibility over which specific creative the user saw and why.<\/p>\n\n\n\n
It isn\u2019t necessarily bad news, but it does mean we need to redefine the marketing team\u2019s role. The manual task of creating a hundred versions of a headline no longer makes sense. The strategic work of defining messages, unique value propositions and guardrails for the model takes on greater importance. In an agency like ours, this means less time spent in the dashboard and more time refining hypotheses, briefs and attribution models.<\/p>\n\n\n\n
In the accounts we manage, we\u2019ve started doing three things. We\u2019re auditing Merchant Center feeds to clean up attributes and get the product listings ready for Gemini to use effectively. We are reworking the Performance Max assets to deliver more message variations rather than a hundred versions of the same headline. And we are reviewing each client\u2019s website for inconsistencies or gaps that a Gemini agent could amplify if it ends up being trained on that information.<\/p>\n\n\n\n
These changes may not be particularly glamorous, but they are the ones that will make all the difference when these formats eventually arrive in Spain. Agency advertising rewards those who come prepared and penalises those who wing it.<\/p>\n\n\n\n
If you\u2019re still unsure how to apply all this to your business, get in touch. We can sit down together to review your Google Ads account and, above all, what lies behind it: your product feed, your website and your analytics. Because that\u2019s where the real game is played from now on.<\/p>","protected":false},"excerpt":{"rendered":"
Google Marketing Live 2026 has unveiled Ask Advisor, Conversational Discovery Ads, Business Agent for Leads and AI-powered Shopping Ads. We take a look at what this means for your campaigns and how you can prepare.<\/p>","protected":false},"author":2,"featured_media":2079,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[103,14],"tags":[],"class_list":["post-2098","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-inteligencia-artificial-2","category-ppc-management"],"yoast_head":"\n