SEO, SEM and AI News | September 2026

The latest SEO, SEM and AI news from the final week of September 2026 brings several changes worth following closely. Google has launched a new spam update, Search Console is beginning to show multimodal search data, AI Max continues to absorb legacy Google Ads functionality, and OpenAI is accelerating the development of ChatGPT Ads with conversational formats and commercial integrations.

This week’s SEO, SEM and AI news also includes developments around first-party data, agentic commerce, YouTube, Gmail, Google Maps and Microsoft Advertising. Not all of them require immediate action. Some may already affect accounts and projects, while others are more strategic and help us understand where digital marketing is heading in 2026.

That is precisely the approach I want to maintain in this weekly SEO, SEM and AI news series: separating what has changed, why it matters and what actually deserves action from all the noise surrounding SEO, digital advertising and artificial intelligence.

The broader direction is becoming increasingly clear. Search engines and advertising platforms are using more artificial intelligence to interpret queries, select ads, adapt creative assets, choose landing pages and accompany users throughout a purchase journey. At the same time, new ways are emerging to measure visual searches, provide context to algorithms and connect advertising, CRM, ecommerce and AI agents.

This development fits with the broader transition I have discussed when analysing the evolution from SEO towards GEO and AI-driven search: we are not suddenly replacing one discipline with another. We are adding new layers to an infrastructure that still requires strong content, reliable data, measurement, authority and a clear strategy.

SEO, SEM and AI news of the week at a glance

Before looking at each development in detail, this table summarises the fifteen SEO, SEM and AI news items that I consider most relevant from this week’s briefing.

Update	Area	Why it matters
September 2026 Spam Update	SEO	It may cause ranking fluctuations while the rollout continues.
Multimodal search in Search Console	SEO / Analytics	It allows traffic from Lens, Circle to Search and image-based searches to be measured.
Migration of features towards AI Max	Google Ads	ACA and campaign-level broad match are entering the automatic transition.
AI Brief and new AI Max reports	Google Ads	More context for AI and more transparency around queries, ads and landing pages.
IP and timestamp in Customer Match	Google Ads	New signals for improving matching, although with restrictions in Europe.
Greater emphasis on first-party data	Measurement	Google continues to stress the importance of proprietary data for feeding its systems.
Sponsored Agents	ChatGPT Ads	An ad can become a conversation with an agent representing the company.
ChatGPT Ads available in Spain	Advertising	Self-service access now includes companies established in Spain.
HubSpot and Shopify integrations	ChatGPT Ads	OpenAI is connecting advertising with CRM and ecommerce.
Universal Commerce Protocol	Ecommerce / AI	Google is developing infrastructure for carts, checkout and agentic commerce.
Business Agent in YouTube Ads	Demand Gen	Users can ask questions about products directly alongside the ad.
One-click journeys from Shorts and Gmail	Demand Gen	Fewer steps between the creative and the landing page.
Promoted Pins in Google Maps	Local advertising	It expands the possibilities for capturing local demand through Maps.
AI Max in Microsoft Advertising	SEM	Microsoft is following the move towards more automated and contextual Search campaigns.
More caution around “GEO recipes”	GEO	Correlations and controlled tests are not the same as universal visibility factors.

Latest SEO news: Spam Update and multimodal search

Within this week’s SEO, SEM and AI news, the two most important changes for organic search come directly from Google Search: a new spam update and an improvement to Search Console that allows multimodal searches to be analysed.

Google launches the September 2026 Spam Update

The most immediate SEO development began on 24 September, when Google started rolling out the September 2026 Spam Update. The rollout affects Google’s ranking systems and, according to the information published by Google, may take approximately two weeks to complete.

At the time of writing, the rollout is still active. This means that any ranking movement recorded during these days should be interpreted with caution. A URL may fall, recover some ground or move again before Google confirms that the update has finished.

Why it is better not to react too quickly

One of the usual temptations when an update appears is to explain every movement immediately. If a page loses rankings on the same day, we look for a cause. If another page rises, we tend to interpret the movement as validation of what we were doing.

The problem is that during a rollout we still do not have a stable picture.

Google has also not presented this update as a new independent list of rules. It remains more useful to analyse the project from a broader perspective: content quality and usefulness, patterns designed mainly to manipulate search results, authority, artificial links, scaled pages with insufficient value or any other practice that may conflict with Google’s spam policies.

The rollout status can be monitored directly through the Google Search Status Dashboard.

What to analyse if a website loses rankings

If a project records significant changes, I would preserve the data and analyse patterns by directory, page type, query and competitor.

What I would avoid is making large-scale changes to a website simply because it has moved during the first few days. In an SEO audit, the context of an algorithm update matters, but it never replaces a complete diagnosis.

Search Console starts measuring multimodal searches

The second major piece of SEO, SEM and AI news this week is particularly interesting because it gives us access to data that until now was difficult to isolate.

Google has added information about multimodal web search to Search Console. This category includes queries made through Lens, Circle to Search, images uploaded directly to Google Search and the Chrome functionality that allows users to search with an image.

What changes inside Search Console

The change appears in performance reports through a new multimodal search filter. Google has also indicated that this data is integrated both into traditional search performance reporting and into reports related to generative AI experiences.

This matters because visual search can no longer be understood as something that happens exclusively inside Google Images.

A user can photograph an object, highlight part of an image or combine visual and textual information to begin a search. From there, the user may still end up on a website in much the same way as with a conventional query.

Google explains the new filter in its announcement about multimodal search in Search Console.

Which sectors may benefit most from visual search

For ecommerce, fashion, interior design, hospitality, tourism, automotive businesses or any sector where visual content plays an important role, I would already begin saving a baseline.

Even if the volume is still small today, having a reference point will make it possible to see in a few months whether this type of discovery is growing and which pages or images are involved.

It is also another reason not to reduce image SEO to filling in an ALT attribute. We need correctly served files, semantic context, useful pages, images that genuinely provide information and a clear relationship between the visual element and the content around it.

SEM and Google Ads news: AI Max gains more importance

The advertising section of this week’s SEO, SEM and AI news is particularly busy. Google continues moving functionality towards AI Max, is expanding AI Brief, preparing new reports and strengthening the use of first-party data.

Google Ads continues moving legacy functionality towards AI Max

In SEM, the most significant development once again revolves around AI Max for Search campaigns.

Google initially announced a faster transition for Dynamic Search Ads and other legacy functionality. It later changed the timeline. The automatic DSA upgrade has been delayed until February 2027, while campaigns using Automatically Created Assets and campaign-level broad match continue entering the automatic migration towards AI Max from September 2026.

The difference matters because for several months September was discussed as though all of these features were changing at the same time. Google later updated its announcement about the migration of Dynamic Search Ads to AI Max to reflect the revised timetable.

What AI Max really changes

AI Max tries to expand campaign reach by using more context than a simple keyword list. It can use keywords, ads, website content and other signals to identify relevant queries, adapt copy and select landing pages.

Automation may uncover opportunities we had not anticipated. The risk begins when we stop understanding which query was used, what message the user saw or why the user ended up on a particular URL.

AI Brief arrives in Spanish

Google is expanding the closed beta of AI Brief to several languages, including Spanish.

The idea is to allow advertisers to explain their business, audience and key messages in natural language. Instead of providing only keywords, assets and settings, we can add a layer of context that is closer to a strategic brief.

The development makes sense to me. Two companies may use very similar keywords while requiring completely different messages because of pricing, positioning, audience, margins, service model or value proposition.

If AI Max is going to make more decisions, it makes sense that we should be able to explain more clearly which business it is trying to represent.

Google prepares query, creative and final URL reporting

The other half of the announcement tries to recover some of the transparency that can be lost through increasing automation.

Google has announced a unified view designed to explain which search term triggered the ad, which creative assets the user received and which landing page they eventually reached.

AI Brief and the new reporting experience were presented together in Google’s update on new AI Max features.

For me, this second point is even more important than having access to more automated generation.

If a system has the freedom to expand queries, adapt creatives and select URLs, we need tools that let us audit those decisions. Automation should not mean giving up our ability to understand what is happening.

Customer Match adds IP and timestamp signals

Google Ads has also expanded the signals that can be used in Customer Match. Data Manager documentation now allows the IP address captured during an interaction, together with the associated timestamps, to be provided to improve audience matching.

The limitation for Spain and the European Economic Area

There are two important details. First, these values are sent without hashing. Second, Google does not support this form of matching for end users located in the European Economic Area, the United Kingdom or Switzerland.

The technical documentation is available in Google’s guide to sending audience members through the Data Manager API.

For a company focused on Spain, therefore, the immediate practical impact is limited. Even so, the development confirms the increasing importance Google places on proprietary data and on the quality of the signals advertisers provide.

First-party data becomes even more important

This trend becomes clearer when we connect it with another item from September’s SEO, SEM and AI news. Google is strengthening its measurement and first-party data activation infrastructure through Data Manager, Google tag and new tools designed to evaluate data strength.

The company argues that a strong base of proprietary signals is one of the conditions required for its AI systems to optimise towards business objectives.

Google develops this idea further in its update about data and measurement tools.

Not every conversion has the same value

From a practical perspective, this again highlights something that many accounts still handle poorly: not every conversion has the same value.

A submitted form may represent a genuine prospect, a supplier trying to sell something to us, an enquiry with no budget or a highly valuable commercial opportunity.

If the algorithm receives only the signal “form submitted”, it will try to find more people capable of carrying out that action. The better we connect CRM information, opportunities, sales and offline conversions, the more real business information we can send back to the platform.

Advertising automation depends less and less on manually configuring every decision and increasingly on the quality of the data used to feed the system.

Artificial intelligence news: ChatGPT Ads grows as an advertising platform

Several OpenAI developments also stand out among this week’s SEO, SEM and AI news. ChatGPT Ads is still developing, but it is starting to look considerably broader than a simple advertising format inserted into a conversation.

Sponsored Agents: from an ad to a conversation with the company

OpenAI is testing Sponsored Agents, a format that allows users to begin a conversation with a company-sponsored agent after clicking an ad.

The difference compared with traditional advertising is significant. The ad does not necessarily have to send the user straight to a landing page.

It can first open a conversational layer in which the person asks about specifications, compatibility, differences between products or any other question related to the offer.

The test is still limited, but conceptually it points towards advertising where creative, recommendation and commercial assistance begin to overlap.

OpenAI introduced Sponsored Agents as part of its update on new AI-powered advertising experiences.

This development also connects with the way users increasingly interact with generative assistants. In my article about how ChatGPT searches the web, I explain the difference between consulting a list of results and maintaining a conversation in which the system can search, reorganise information and continue answering questions about the same problem.

ChatGPT Ads now offers self-service access in Spain

The second development is much more immediate for our market.

Spain is now listed among the countries where ChatGPT Ads Manager self-service access is available to legal entities established there.

The current list can be reviewed on the official ChatGPT Ads Manager availability page.

How to measure an advertising channel that is still new

This does not mean that ChatGPT Ads can be judged using the same historical benchmarks or level of maturity as Google Ads or Meta.

Precisely because the ecosystem is still developing, early tests should be treated as controlled experiments and measured against business criteria.

CTR, CPC or volume may be useful diagnostic metrics. What will ultimately determine whether the channel deserves budget is the quality of opportunities, sales and attributable revenue.

HubSpot and Shopify become strategic integrations

OpenAI has also announced HubSpot as its first CRM partner and Shopify as its first ecommerce partner within this new advertising ecosystem.

The logic is quite clear. If ChatGPT Ads wants to become a relevant platform for businesses, it needs to connect with the systems where commercial data, products and customers already live.

CRM integration brings advertising closer to actual commercial outcomes. Ecommerce integration makes it easier to use catalogue information and manage campaigns within a workflow that is closer to the real business.

Once again, we see the same trend as with Google: advertising, proprietary data, catalogues and automation are beginning to form a much more connected system.

Ecommerce and AI: Google moves towards agentic commerce

Another strategically important item in this week’s SEO, SEM and AI news concerns ecommerce and artificial intelligence agents.

Google develops Universal Commerce Protocol

Google continues to develop the Universal Commerce Protocol, or UCP, an infrastructure designed to allow merchants, agents and Google services to exchange structured information throughout a purchase process.

UCP includes functions such as identity linking, cart creation and transfer, checkout, order management, discounts and loyalty. AI Mode in Search and Gemini are among the surfaces involved.

The Universal Commerce Protocol documentation shows that this is not merely a conceptual idea. Google already provides technical documentation for cart functionality and native checkout.

What agentic commerce means for ecommerce businesses

This has interesting implications for online stores.

For years, we have structured catalogues so that Google could crawl pages, understand products and work with Merchant Center. A new objective is now appearing: allowing an agent to use that information inside a shopping experience.

A product with inconsistent attributes, unreliable availability, poorly defined variants or incomplete commercial data does not only create an SEO or Shopping problem. It may also make it harder for an automated system to understand what is being sold and how it can be purchased.

That is why feeds, structured data and catalogue quality are beginning to form a shared infrastructure across SEO, advertising and agentic commerce.

YouTube, Gmail and Google Maps add new advertising functionality

Google has also grouped several developments under its September 2026 Demand Gen Drop. They all share a similar goal: reducing the distance between discovery and action.

Business Agent arrives in YouTube Ads

Business Agent for YouTube Ads allows users to interact with a conversational AI alongside certain video ads connected to product feeds.

A person can see a product and ask questions without immediately leaving the video context. Once again, an intermediate layer appears between advertising exposure and the website visit.

Conceptually, it is similar to Sponsored Agents, although within Google’s ecosystem: advertising begins to incorporate the ability to answer questions before the user moves towards conversion.

YouTube Shorts and Gmail reduce the journey to a single click

Google is also introducing one-click experiences from certain YouTube Shorts and Gmail display ads directly to the landing page.

In Shorts, the change is particularly related to full-screen 9:16 vertical creatives. It may seem like a minor interface improvement, but reducing unnecessary steps remains one of the simplest ways to reduce friction.

Naturally, reaching a page more quickly does not fix a weak proposition. If the landing page does not respond to user intent, we will simply be sending traffic through fewer steps towards an experience that still fails to convert.

Promoted Pins expand advertising in Google Maps

The third development affects Google Maps. Google is expanding Promoted Pins through Affiliate Location Extensions for certain retail networks and dealerships.

Promoted pins can appear while the user navigates, searches or explores Maps.

It is another reminder that local visibility cannot be analysed solely through the organic ranking of a business profile.

Google Business Profile, Search, Maps and Google Ads form an increasingly connected ecosystem. When working on local SEO and Google Business visibility, I therefore try to distinguish clearly between visibility generated through organic positioning and visibility created through advertising investment.

The three features are included in Google’s official announcement about the September 2026 Demand Gen Drop.

SEM news for 2026: Microsoft Advertising also bets on AI Max

This week’s SEO, SEM and AI news does not end with Google and OpenAI. Microsoft Advertising is also accelerating automation inside its Search campaigns.

AI Max arrives in Microsoft Advertising

Microsoft Advertising has made AI Max for Search available, an optional suite that combines search term matching, text personalisation and final URL expansion.

The system uses keywords, ads, landing pages and other contextual signals to try to identify relevant searches that were not explicitly included in the keyword list, generate or adapt messaging and connect the user with an appropriate landing page.

Microsoft retains brand controls, exclusions and reporting designed to allow advertisers to test expansion without completely losing oversight.

The company explains the system in its presentation of AI Max for Search.

Google and Microsoft are moving in the same direction

Beyond comparing one platform with another, I find it significant that both are moving in a very similar direction.

Search is becoming less dependent on a closed relationship between keyword, ad and URL and increasingly uses context and models capable of interpreting intent.

GEO news in 2026: be careful with universal optimisation recipes

One of the SEO, SEM and AI news topics that requires the most reflection is not the launch of a specific tool, but the growing number of studies and supposed optimisation recipes appearing around GEO.

Every few weeks we see a study, correlation or test trying to identify what makes a brand or page appear more frequently inside generative responses.

That research is necessary. The problem begins when a specific result is quickly transformed into a universal checklist.

Correlation does not mean causation

A variable being correlated with higher visibility does not automatically prove causation.

Likewise, a technique increasing visibility in an experimental environment does not prove that it will cause a source to be discovered, retrieved and selected from millions of documents under real-world conditions.

The 2026 academic review of GEO is useful precisely because it separates stages that are sometimes mixed together: discovery, retrieval, selection, context, citation and presentation of the response.

GEO needs longitudinal measurement

We can learn from correlations involving mentions, structure, citations or authority, but we should not turn them into the equivalent of a new keyword-density formula.

In my analysis of the best APIs for SEO and GEO, I already explained that measuring generative visibility requires tracking prompts, mentions, competitors and sources over time.

The objective should be to detect patterns, not to chase a single answer as though it were a fixed ranking position.

What I would do this week with these SEO, SEM and AI updates

Of all the SEO, SEM and AI news analysed above, there are five actions that I consider concrete enough to incorporate into this week’s work.

1. I would not draw final conclusions while the Spam Update is still active

I would save the data, compare pages and queries and review any unusual movement.

But I would wait for the rollout to progress before making large-scale changes based solely on a recent rise or fall.

2. I would create a multimodal search baseline in Search Console

I would check whether the property is already receiving data and export an initial reference.

For visually driven projects, this may become an interesting metric, and the sooner we have a baseline, the easier it will be to interpret its evolution.

3. I would audit campaigns using ACA or campaign-level broad match

I would review which accounts may be entering the AI Max migration, what settings remain active and how much control we still retain.

I would also review exclusions, URLs and the signals being used.

4. I would test AI Brief and the new reporting when they become available

AI Brief remains in closed beta, so it will not yet appear in every account.

When it does, I would not simply fill it in. I would prepare a genuine business brief and observe whether it changes the quality of terms, messaging and traffic.

The new query, creative and final URL report also looks particularly useful because it should help us understand whether automation is making decisions that actually make sense.

5. I would review feeds and structured data with AI agents in mind

In ecommerce, I would begin treating catalogue information, availability, attributes, variants and structured data as a shared infrastructure across SEO, Shopping and agentic systems.

AI-driven commerce is still evolving, but waiting until it becomes mainstream before organising the data would make little sense.

What this week’s SEO, SEM and AI news tells us as a whole

If we analyse these SEO, SEM and AI news items individually, they can look like completely different issues: algorithms, campaigns, ecommerce, agents, data or GEO.

But there is a clear thread connecting almost all of them.

Platforms are increasing the number of decisions they can make automatically.

Search interprets increasingly complex queries. Google Ads decides which searches may be relevant and which URL to use. ChatGPT can turn an ad into a conversation. YouTube is beginning to place agents around products. Ecommerce systems are preparing for AI to participate directly in the purchasing process.

This moves part of the work away from manual execution and towards information quality, data, supervision and strategy.

Javier Carmona Benítez’s opinion on SEO, SEM and AI news

I have been working professionally in digital marketing and SEO since 2013, and one of the things that stands out most to me in this week’s SEO, SEM and AI news is how difficult it is becoming to separate SEO, SEM, data, ecommerce and artificial intelligence completely.

For years, the boundaries were much clearer.

In SEO, we worked on content, architecture, links and rankings. In Google Ads, we configured campaigns, search terms, ads and bids. Ecommerce teams managed their catalogues and Merchant Center. Analytics tried to measure what happened afterwards.

Now almost all of these elements are beginning to connect.

Search Console is adding multimodal search because a query no longer has to begin with text. AI Max uses information from the website and the advertiser to decide how to respond to a search. First-party data helps advertising systems distinguish which users create value. ChatGPT Ads is experimenting with sponsored agents capable of holding conversations. Google is developing protocols that allow AI to work with shopping carts and checkout processes.

I do not believe this means that the previous disciplines are disappearing.

I think something very similar is happening to what we have already seen throughout the evolution of SEO: previous layers continue to exist, but they become part of more complex systems.

Keywords are still useful. Links still matter. Landing pages still need to convert. A feed still needs to be well built. A campaign still needs to respond to a business objective. And a company still needs sufficiently good data to know whether it is gaining customers or simply accumulating events inside a platform.

Automation does not remove the need for judgement

The difference is that an increasing share of execution can now be handled by a machine.

If AI Max can expand queries, adapt creatives and choose URLs, the value of the professional is not in trying to compete with the algorithm by manually producing every possible combination.

It lies in defining the objective correctly, providing high-quality information, establishing boundaries, interpreting the data and identifying when automation is working and when it is making poor decisions.

The same applies to GEO

It is tempting to search for a simple new formula: number of mentions, paragraph format, quantity of citations or any other signal that can be transformed into a checklist.

But a generative response depends on discovery, retrieval, context, sources, model, query and other elements that we cannot fully control.

That is why I prefer to work with principles that remain useful even when the platform changes: a technically accessible website, content containing real information, first-hand experience, distributed authority, a coherent entity and sufficient measurement to understand where we are appearing.

First-party data is also becoming a competitive advantage

I also believe that this week’s SEO, SEM and AI news reinforces an issue that will become increasingly important: first-party data and structured information are part of marketing, not merely a technical concern.

A company that knows which contacts become customers can train its campaigns more effectively than one that only measures submitted forms.

An ecommerce business with a clean catalogue can perform better in Shopping and, progressively, in agentic experiences.

A brand that appears consistently across different sources provides more information to the systems trying to understand who it is and what it does.

SEO, SEM and GEO are becoming part of the same digital strategy

In that context, SEO, SEM and GEO do not compete with each other. They need each other.

SEO helps make our information accessible, understandable and relevant. SEM allows us to capture and experiment with demand in a controllable way. GEO extends the objective into environments where an AI system may retrieve, compare and use that information before a click even takes place.

As an SEO consultant, I work precisely with this combined approach: maintaining a solid foundation, measuring what actually generates business and adapting the strategy when new technology changes the way people search, compare and make decisions.

You can also review my services as a GEO consultant or find me on Google Maps to see my profile, location and client reviews.

Of all this week’s SEO, SEM and AI news, there is one idea I would probably keep above all the others: artificial intelligence is increasing the ability of platforms to make decisions, but that makes the quality of the data, content and professional judgement used to supervise them even more important.

Technology will continue to change and within a few weeks we will have new features, names and models. The underlying work remains fairly similar: making sure that when someone needs information, a product or a service related to our business, the platforms have enough signals to consider us a relevant answer.