OpenAI has begun rolling out GPT-Live, a new generation of voice models that power ChatGPT Voice. During a conversation, GPT-Live can hand a question to its latest frontier model, such as GPT-5.5, for heavier reasoning or a web search.

What Launched

GPT-Live-1 becomes the default model powering ChatGPT Voice for Go, Plus, and Pro users, and GPT-Live-1 mini becomes the default for Free users.

OpenAI describes the models as full-duplex, meaning they can take in audio and produce speech at the same time, interrupt less, and wait when a user pauses.

Search and Visual Answers Move Into the Conversation

When a spoken question needs deeper reasoning or current information, GPT-Live can hand it to a frontier model, GPT-5.5 at launch, and bring the answer back into the conversation.

ChatGPT Voice can also show visual cards for topics like weather, stocks, and sports, and OpenAI said Voice continues to support search, memory, images, and file uploads. Users can set a reasoning level, with Instant for fast replies and Medium or High for more involved questions. OpenAI said Instant replies and the mini model run on GPT-5.5 Instant, while Medium and High use GPT-5.5 Thinking.

How OpenAI Measured It

In OpenAI’s own head-to-head evaluations, GPT-Live-1 and GPT-Live-1 mini were preferred over its Advanced Voice Mode in conversations lasting five to ten minutes. OpenAI said the evaluations looked at overall preference, turn-taking, interruptions, conversational flow, and how natural the interactions felt. The company said more than 150 million people talk to ChatGPT each week using features like Voice and Dictation.

What It Doesn’t Do Yet

At launch, GPT-Live does not support voice with video or screen sharing in ChatGPT. OpenAI said it is working to add those capabilities, and that its earlier Standard and Advanced Voice Modes stay available where video and screen sharing are supported.

Why This Matters

Now, when you ask a question aloud, you might see an answer directly in the voice flow, sometimes accompanied by a visual card on the screen. This provides ChatGPT with another way to present an answer without visiting a source site. Since the reasoning and search process happens behind the scenes on GPT-5.5, what shows up there depends on how the model retrieves information and cites its sources.

Looking Ahead

OpenAI said the rollout is beginning globally, and that video and screen sharing are not in this release but are being worked on.

OpenAI’s post doesn’t say how GPT-Live handles citations when it answers a spoken question from a GPT-5.5 web search. ChatGPT’s text answers show source links next to the response. Whether a spoken answer names its sources, shows them on screen, or leaves them out is the detail to watch. That’s what decides whether a search inside a voice conversation can still send a reader to your site.

Featured Image: OpenAI

https://www.searchenginejournal.com/openai-gpt-live-brings-search-into-chatgpt-voice/581773/

Multi-channel marketing tends to sound more complicated than it needs to be.

But we live in a world of “more”: more platforms, more campaigns, more assets, and more reporting to explain. The work expands, but the strategy behind each channel is not always clearly defined.

Especially when you’re asked to do more at the speed of light, your multi-channel strategy can easily end up not looking like a strategy at all.

That’s usually where the bigger question shows up: Are these channels actually working together, or are they being measured like separate programs?

That was the focus of Session 3 of SEJ Live, where I joined Shaun Bruno from CallRail to talk about how to build a multi-channel growth strategy that actually converts.

The full session is available to watch on demand here: SEJ Live On Demand.

During the session, we talked through how marketers can connect channels across the customer journey, where AI can help, and what still needs human judgment. We also covered attribution, creative workflows, and how to set expectations before a campaign launches.

This recap covers some of the main takeaways, while the full recording goes deeper into the channel playbooks and audience Q&A.

1. Give Each Channel A Clear Job

One of the biggest issues I see with multi-channel strategies is unclear expectations.

A business may have the right channels in place, but each one is being judged against the same goal. That usually means every campaign is expected to drive immediate conversions, regardless of where it sits in the customer journey.

The problem? That simply doesn’t reflect how people actually make decisions.

Someone might discover a brand through video, compare options through search, look for proof on social or Reddit, and convert later through email or branded search. As consumers, we do this all the time. We gather information in one place, validate it somewhere else, and take action when the timing or offer makes sense.

The challenge is that many teams still manage channels like separate programs.

Paid search and social have their own report. Email, organic, webinars, and video often get evaluated separately, too.

That setup makes it easy to ask which channel “won” the conversion. It makes it harder to understand how the channels worked together.

Before judging performance, marketers need to define the role of each channel.

Some channels are better at creating demand by introducing a problem or product category. Others are better at building trust by answering questions, addressing objections, or showing proof. Lower-funnel channels are usually stronger at capturing intent once someone is closer to making a decision.

Your goal at this step is to understand what each channel should contribute, then measure it against that role. Once that’s clear, the performance conversation becomes a lot more useful.

2. Measure Awareness By The Right Signals

Awareness is one of the easiest places for a multi-channel strategy to get misunderstood.

A brand says it needs more awareness, then ends up measuring those campaigns against bottom-funnel conversion goals. The marketer may be trying to build future demand, while the business is asking why the cost-per-acquisition does not look like branded search.

That gap needs to be addressed before the campaign launches.

If you are running YouTube, Meta, TikTok, CTV, audio, creator, or influencer campaigns, those channels may be reaching people who were not already looking for your brand. They can introduce the problem, build familiarity, and give people a reason to come back later.

Immediate CPA won’t show all of that.

For upper-funnel campaigns, I would look at signals such as branded search lift, returning visitors, direct traffic trends, video completion rates, first-party audience growth, assisted conversions, and downstream performance in other channels.

Those metrics can help show whether the campaign is creating more awareness and improving the path for other channels.

This is also where expectation-setting matters.

Before a campaign goes live, marketers should be clear about how it will be evaluated. That conversation may need to happen with leadership, finance, clients, or managers.

For example, if you’re launching a YouTube campaign, define what success looks like. Explain which metrics matter, which metrics are directional, and which metrics should not be used as the main decision point.

Without that alignment, upper-funnel campaigns are often the first to get cut. Then, a few months later, the same business may wonder why branded demand is not growing.

3. Use Mid-Funnel To Reduce Uncertainty

Mid-funnel is where a lot of brands start to lose people.

At this point, someone has already shown some level of interest. They may have watched a video, visited a product page, downloaded a guide, engaged with social content, or searched for comparison terms.

The next message should not be the same one that introduced them to the brand.

This is where the message needs to shift from problem awareness to proof and differentiation. If someone watched an awareness video, they may need more product education. If they visited a product page, they may need reviews, testimonials, or answers to common objections. If they downloaded a guide, they may need a clearer connection between the educational content and the solution.

The goal is to answer the questions that are keeping someone from moving forward.

Meta can support this with objection-focused creative, carousel ads, or video retargeting. Demand Gen and YouTube can help reinforce product education in a more visual format. LinkedIn can work well for case studies, business proof, and credibility in B2B. Email can help sequence the message once someone has shared their information.

I also think marketers should be careful about treating mid-funnel as only retargeting.

Retargeting can play a role, but mid-funnel can also include non-brand search, webinars, educational content, comparison pages, nurture sequences, and other touchpoints that help someone build confidence.

The question I like to ask is: What is preventing this person from taking the next step?

Maybe they don’t trust the brand yet. Maybe they don’t understand the difference between your product and another option. Maybe they need to see proof from someone who had the same problem.

Once you know what’s causing hesitation, the message sequencing becomes easier to build.

4. Keep Last-Click Attribution In Context

Lower-funnel campaigns often look like the best performers because they sit closest to the conversion.

Paid search, branded search, and retargeting tend to get the lead, the sale, and the credit. But in many cases, those channels are capturing demand that another channel helped create.

This is where last-click attribution can give an incomplete view of performance. It rewards the final touchpoint, even when earlier touchpoints played a meaningful role in the decision.

Last-click data can still be useful, especially for understanding which channels are driving immediate action. It just shouldn’t be the only way performance is evaluated.

In the session, Shaun talked about looking at attribution from multiple angles. That can include last click, balanced attribution models, total influence touchpoints, and view-through data for upper-funnel channels.

That type of reporting gives marketers more context when deciding where to invest.

A B2B company with a long sales cycle should not evaluate every channel the same way as a direct-to-consumer brand with a short buying window. The right approach depends on the business, the sales cycle, and the decisions the data needs to support.

Since attribution will never capture every touchpoint perfectly, the reporting should help marketers make better budget decisions with the data they do have.

Marketers still need to report numbers leadership can understand. Those numbers just need context, especially when channels are designed to play different roles.

5. Use AI To Improve Workflow

AI was a big part of the session because it is already changing how many marketing teams work.

For me, the more useful place to start is workflow. Where can AI reduce repetitive work, speed up analysis, or help the team get to a stronger starting point?

AI can help summarize reviews, organize customer research, cluster search queries, generate headline variations, build creative starting points, resize assets, flag anomalies, and pull themes from performance data.

Those use cases can save real time, especially for smaller teams.

AI doesn’t automatically know your positioning, margins, sales feedback, customer objections, compliance needs, or internal priorities. Those details have to be part of the process if you want the output to be useful.

That’s why I would look at the foundation before bringing AI into every part of the workflow.

• Is tracking clean?
• Are audiences clearly defined?
• Do you know the difference between the buyer, the user, and the influencer?
• Is the messaging clear?
• Do the landing pages answer the right questions?
• Does each channel have a defined role?

If those pieces are unclear, AI can help you move faster in the wrong direction. But, if those pieces are in place, AI can help the team move faster without handing over the strategy.

Build A Strategy That Can Adapt

Multi-channel growth gets a lot easier to evaluate when each channel has a clear role.

That doesn’t mean every path will be easy to measure. It also doesn’t mean every channel will show its value in the same report. But it does give marketers a more useful way to plan, test, and explain performance.

For me, that is where a lot of the work should happen next.

Look at the channels you are already using. Ask what each one is supposed to contribute. Then look at whether the message, campaign objective, landing page, and measurement approach support that role.

If the answer is unclear, adding another channel probably will not fix it.

AI can help improve parts of the workflow, especially around research, creative, reporting, and pattern recognition. But the strategy still needs human judgment. Marketers still need to define the customer, the message, the offer, and the signals that matter.

For anyone who wants the full discussion, including the channel playbooks, AI workflow examples, and audience Q&A, the complete session is available to watch on demand here: Watch SEJ Live On Demand.

SEJ Pro is also where the conversation continues beyond the live sessions. It gives marketers a place to ask follow-up questions, talk through what they are testing, and learn from experts and peers working through similar challenges.

A lot of teams are being asked to do more with less right now. Having a place to talk through what is working, what still feels unclear, and what deserves more testing can make those decisions feel a little less isolated.

Featured Image: Master1305/Shutterstock

https://www.searchenginejournal.com/5-ways-to-make-your-marketing-channels-work-together/581005/

For more than two decades, digital strategy has revolved around a deceptively simple objective: Drive people to webpages. Search engines rewarded documents. Analytics rewarded pageviews. Marketing rewarded engagement. As organizations matured, they invested heavily in designing increasingly sophisticated digital experiences that guided customers through carefully orchestrated buying journeys. Information was intentionally distributed across dozens, sometimes hundreds, of interconnected pages, each optimized for a different stage of consideration.

Consider how a company such as Ford presents the F-150, one of the best-selling vehicles in America. Rather than offering a single comprehensive representation of the vehicle, Ford brilliantly guides prospective buyers through an emotional journey spread across seven distinct viewports. The homepage establishes the lifestyle. Model pages introduce trim levels. Interactive configurators allow customers to visualize ownership. Feature pages explain towing capacity, off-road performance, and technology packages. Galleries reinforce the brand’s identity, while technical specifications are located deeper within the site, alongside regional offers and financing options.

For people, this architecture works remarkably well. Every page serves a purpose. Every interaction builds confidence. Every transition moves the customer toward a purchase decision. It is an outstanding human experience. For AI, however, the same architecture introduces friction.

The Quiet Crisis Of AI Disintermediation

The AI labs frequently tell enterprise leaders that their large language models (LLMs) are smart enough to crawl any messy web architecture, synthesize the data, and deliver accurate answers regardless of how that information is organized. That message oversimplifies reality and how AI retrieval actually works.

When data is deliberately fractured across multiple pages to serve human emotions, the AI’s synthesis engine breaks. Because the machine lacks an emotional context window, it searches for a high-density, low-latency semantic payload. When it cannot find that payload natively on an official corporate domain, it looks elsewhere. It then assembles the most complete answer it can from whichever sources are easiest to retrieve, reconcile, and trust. The consequences are already visible.

A straightforward query such as [ford f-150 Raptor gas mileage] produces a Google AI Overview that draws information from Reddit discussions, automotive publishers, and a local dealership rather than Ford itself.

Ford already has the answer to nearly every conceivable question. The issue isn’t that the information doesn’t exist. The issue is that Google found it easier to assemble an answer from Reddit, an automotive publisher, and a dealership than from Ford itself. When that happens, the discussion is no longer about rankings or citations. It is about who controls the authoritative representation of your brand.

This is no longer simply an SEO problem. It is a content governance problem.

The issue is that AI has simply exposed a structural weakness that has existed for years. Enterprises organized their digital presence around webpages because search rewarded webpages. In many ways, search became the detour. Organizations optimized for ranking documents and triggering an emotional reaction rather than organizing knowledge. That approach worked because search engines retrieved pages. AI assistants attempt to synthesize a coherent representation of the organization. In doing so, they expose every inconsistency, every missing relationship, and every gap in the underlying architecture.

The organizations struggling today are rarely missing information. They possess enormous knowledge of their products, services, policies, and expertise. The problem is that the knowledge has been fragmented across webpages, content management systems, product databases, marketing campaigns, PDFs, support portals, and countless disconnected repositories. Humans can navigate those silos. Machines increasingly cannot.

AI did not create this problem. It simply made it impossible to ignore.

Brand Sovereignty Becomes An Executive Responsibility

Years ago, I had the opportunity to consult for Dell, where Michael Dell demonstrated an approach to digital leadership that feels even more relevant today than it did then. He regularly tested both Google Search and Dell’s internal search experience himself, not because he wanted to micromanage marketing or technology teams, but because he understood something many executives overlooked: the interface through which customers discover your products ultimately shapes how they perceive your company.

If he or a customer searched for a product and failed to find the right answer, Michael Dell did not see an isolated technology issue. He saw an organizational failure. That mindset has become even more important in the AI era.

I think of this as brand sovereignty: an organization’s ability to remain the authoritative source for information about its own products, services, and expertise, regardless of where those answers are ultimately delivered. For years, digital success was measured by how effectively organizations attracted visitors to their websites. Increasingly, a more important question will be whether AI systems consistently recognize the organization itself as the best source of that information.

This isn’t something marketing, SEO, or technology can solve on their own because none of those teams owns the complete picture. Product information, documentation, customer support, legal policies, and commerce all contribute to how an organization is represented digitally. Reclaiming brand sovereignty, therefore, becomes less about publishing more content and more about organizing organizational knowledge so that those pieces reinforce one another rather than compete.

From Pages To Knowledge

Most organizations didn’t set out to fragment their knowledge. It happened gradually. Every project added another page, another microsite, another content repository, or another system designed to solve a specific business problem. Over time, product information, marketing content, customer support, policies, and commerce evolved independently while the corporate website became responsible for stitching everything together into a coherent customer experience.

That approach worked because the web rewarded navigation. Customers could move between pages, and search engines could retrieve the most relevant document. Neither required organizations to explicitly connect the relationships between their products, services, policies, and expertise.

AI exposes the limitations of that model. Large language models are not attempting to navigate websites in the way people do. They are attempting to understand organizations by reconstructing the relationships between products, services, documentation, policies, locations, expertise, and supporting evidence. Every answer generated by an AI assistant represents an attempt to assemble that understanding from the information available to it. When those relationships remain implicit, distributed across hundreds of webpages, databases, and disconnected repositories, the resulting representation becomes incomplete or inconsistent.

The solution is not publishing more content. It is organizing knowledge differently through a new architectural model.

Rather than treating products, services, documentation, policies, reviews, offers, support resources, and locations as independent publishing assets, organizations should begin managing them as interconnected business objects within a Unified Object Graph. Each object maintains its own identity while explicitly connecting to every related object throughout the enterprise. A product connects to its technical documentation, compatible accessories, warranty information, inventory, customer reviews, dealerships, and service locations. The webpage becomes one expression of those relationships rather than the place where those relationships are created.

One of the questions I hear most often is whether this requires replacing existing systems. In most cases, it doesn’t. Organizations have already invested heavily in product information systems, content management systems, commerce platforms, digital asset management, and customer support tools. Those systems continue to serve important purposes and should remain the systems of record for the information they manage best. The challenge is that none of them represents the organization as a whole.

Instead of trying to consolidate everything into a single platform, organizations should focus on creating a machine-readable knowledge layer that brings those pieces together. Product information, documentation, policies, reviews, marketing content, and commerce data continue to live where they belong, but they are aggregated into a single, machine-readable representation that explicitly describes the entities and relationships across the business.

Once that layer exists, the conversation changes. Publishing to a website, exposing an API, generating structured data, supporting an MCP endpoint, or adopting whatever protocol comes next all become different ways of expressing the same underlying knowledge rather than separate implementation projects.

This is the architectural shift that AI is exposing. For years we managed channels independently and treated the website as the place where everything came together. Increasingly, organizations will manage knowledge centrally while allowing every interface to consume the same authoritative representation. Websites, customer support portals, AI assistants, commerce platforms, and future interfaces all become consumers of the same knowledge rather than maintaining their own versions.

That shift also changes how content is created. Most organizations still separate technical accuracy from marketing language because different teams own different parts of the story. Product Information Management systems manage specifications, creative teams develop messaging, SEO teams research customer language, and customer support documents common questions. Each group adds value, but very little of that knowledge remains connected once it leaves the team that created it.

Consumers, however, do not separate facts from feelings when making decisions. A customer searching for [the safest family SUV], [a truck that feels unstoppable off-road], or [a quiet hotel for remote work] combines objective requirements with subjective expectations in the same question. Increasingly, AI systems are expected to interpret those blended expressions of intent in much the same way.

At Bisan Digital, we call this emotifacts (where feeling and fact are inseparable), and they become valuable to the process because they combine factual product attributes with the emotional language customers naturally use to describe, discover, and ultimately choose products or services. Rather than treating emotional messaging as creative copy layered onto technical specifications, both are treated as part of the same reusable knowledge object.

If marketing positions the Ford Raptor around freedom, confidence, and rugged independence, those ideas should be explicitly connected to the engineering evidence that supports them: suspension travel, approach angles, locking differentials, horsepower, towing capacity, and terrain management systems. The emotional promise and the technical proof reinforce one another because they originate from the same underlying object. The same principle extends well beyond the automotive industry. A luxury hotel should connect its promise of tranquility to room location, sound insulation, wellness amenities, and guest reviews. A healthcare provider should connect claims of clinical expertise to physician credentials, treatment outcomes, published guidelines, and patient education. In each case, trust is strengthened because the emotional narrative and the supporting evidence are inseparable.

This represents the broader transition from digital publishing to knowledge architecture. Machines can infer many things, but they should not be expected to infer the relationships that organizations already know to be true. Increasingly, competitive advantage will belong to the organizations that explicitly declare those relationships, govern them consistently, and make them available across every interface through which customers and intelligent systems engage with the business.

Building For Adaptability Rather Than Standards

Once knowledge becomes independent from presentation, exposing it to both people and machines becomes significantly easier. This is where much of today’s conversation around AI interoperability is focused, and understandably so. New protocols, APIs, and discovery mechanisms are emerging almost monthly as organizations race to determine how AI assistants should access trusted enterprise information.

Emerging standards such as MCP represent an important shift toward explicit machine interfaces. Today’s protocol may be MCP. Tomorrow it may be another widely adopted standard. The objective is not to predict which protocol will win but to organize knowledge so it can be exposed through whichever standards ultimately become dominant.

The same principle applies to commerce. Emerging initiatives such as Google’s Universal Commerce Protocol (UCP) illustrate how structured product knowledge can flow directly into AI-assisted purchasing experiences. Whether UCP becomes the dominant protocol is less important than ensuring the underlying knowledge is structured well enough to participate in whichever transactional ecosystem emerges.

This distinction between architecture and implementation has always mattered, but it has rarely been as visible as it is today. Organizations that continue to treat their website as the primary repository of business knowledge will find themselves repeatedly adapting to new interfaces, new protocols, and new retrieval models. Organizations that instead invest in well-governed, reusable knowledge assets will discover that supporting new delivery mechanisms becomes an incremental engineering exercise rather than a fundamental organizational transformation.

The conversation, therefore, should not begin with MCP, UCP, or any other emerging specification. It should begin with a more fundamental question: Does the organization possess a coherent, authoritative representation of its own knowledge independent of the interfaces through which that knowledge is delivered? Every protocol introduced over the coming decade will simply become another window through which that knowledge can be expressed.

The New Measure Of Digital Success

For much of the web’s history, digital success was measured by a familiar collection of metrics: rankings, website traffic, pageviews, engagement, and conversions. Those measures remain valuable because websites will continue to play an important role in how organizations communicate with customers. They are no longer, however, the only measure of digital effectiveness.

As AI assistants increasingly become intermediaries between organizations and consumers, a new question emerges. When an intelligent system answers a question about your company, your products, or your expertise, does that answer originate from your organization’s knowledge, or from someone else’s interpretation of it? That distinction defines brand sovereignty.

The organizations that succeed during the next decade will not necessarily publish more content than their competitors, nor will they build the most sophisticated websites. They will recognize that digital strategy is no longer centered on documents but on knowledge itself. Their webpages, mobile applications, customer support experiences, AI assistants, commerce platforms, and technologies yet to be invented will all become distinct expressions of the same authoritative foundation.

Search taught organizations how to build better webpages. The AI era is teaching them how to build better knowledge.

The organizations that win the AI era will not be the ones with the most webpages. They will be the ones with the best-organized knowledge.  Your website is no longer your digital asset. Your knowledge is. The website is simply one way of expressing it.

More Resources:

Featured Image: Roman Samborskyi/Shutterstock

https://www.searchenginejournal.com/reclaiming-brand-sovereignty-in-the-ai-era/581161/

I had a conversation with an old friend from my WebmasterWorld Forum days about PR marketing for AI search. The friend had contacted me to hear my thoughts about it. The discussion seemed useful, so I rewrote it into an article.

Digital PR Outreach Because Links Matter Less

The friend I had this conversation with is Alistair White (LinkedIn profile), a search marketing professional based in Australia who has decades of experience.

White asked me:

“I was wondering if you have any thoughts on performance based digital?”

My response was, yes, I have a load of thoughts on the topic. The following is one of them. In the future, I will do a follow-up on more ideas.

I used to do PR outreach slash brand marketing for a B2B starting around 2004. But I scaled it up for another company around 2013 because I saw the writing on the wall that links were already on the decline. So my approach grew out of link building, but my intuition was that links did not matter. It was about putting the company in front of ten thousand, twenty thousand, sixty thousand potential customers and doing it in a way that makes it clear that this company solves the problem that these professionals have.

Strategy: Outreach Directly To Potential Customers

What I did was narrowly focus on a specific demographic that strictly lined up with their target customer. So, one typical customer was the head of IT and IT workers at a large corporation. Another demographic was the department manager. Two different demographics that both needed the same solution. So the campaign was split into two parts, one for each demographic.

It was essentially a PR campaign that was focused on identifying associations and organizations. Virtually every industry has an association of professionals. So, what I did was first target the organizations at the national level. The reason is that once you do a project with the national level, getting similar projects done at the state level was ten times easier. You just show them the national level article that featured the company, and the state-level organizations would almost always say yes.

Once you got that state-level project done, getting to yes with the individual chapters at the regional or county level became ten times easier. Each time I got a project done, it put the client in front of thousands of potential customers. Eventually, everyone knew who my client was and getting projects done became easier.

What were these projects?

• Newsletters
• Organization magazines
• Website articles
• Interviews

Every organization was different. So I would click around and see what they were up to and create the pitch to fit with what they were publishing.

Attribution Is Not Always Possible

This was not about building links, it was about building customers, making money.

And that’s not something that any SEO thirteen years ago would ever consider doing because there isn’t a clear way to track that the client spent X this month and earned Z the next month because of that activity. An SEO would never consider it because there’s no way to directly track the ROI.

You can track some of it with the “how did you hear about us” question. But how will you know if a customer heard of the company because a colleague at a conference who saw the client’s propaganda told them about it?

Not everything can be tracked. That’s why everyone else in SEO did not pursue these opportunities because they were like, where is the link, where is the attribution? Well, now the secret’s out. It’s infinitely adaptable, too.

Both companies that I did this for experienced year-over-year steady growth, and both were eventually acquired and made a lot of money for the founders. And how did I know it worked? Because this is how I promoted some of my websites, by building top-of-mind awareness, the kind that makes people type a domain name into the search box.

Google has algorithms that track things like branded navigation. You can’t build that kind of user behavior with links. You cannot build branded navigation with SEO. And yet, these are things that have been a part of Google’s algorithms since 2004 with the Navboost algorithm and in 2012 with Google’s branded navigation.

Brand Marketing And PR… And SEO?

SEOs have historically been about five to ten years behind the actual algorithm developments at Google. And I get it that ideas that a five year old can understand, like “adding EEAT” to articles, that’s easy to understand. Everyone else is doing it. But you know, everyone who did that knows now it was a grand waste of time.

And it’s not that I am a contrarian. It’s just that most of the time SEOs chase these ephemeral tactics with an SEO hammer, going bang, bang, bang. But it’s becoming clearer now that SEO for AI search is not the nail that building links used to be.

Like, how are you going to encourage people to do a branded search on Google? How are you going to get people to know about your brand in the first place? How are you going to target the office manager that makes the decisions at a company? Well, I shared an idea about that, right?

Those are the kinds of things that are going to trigger a positive ranking factor at Google, and yet none of those are a part of the SEO toolbox.

Back in 2015 I raised the idea that content is king misses the point of being successful online.

I wrote:

“If content is king, how come the top Internet businesses sites are not in the content business? What about Netflix? You think that’s content you are paying a monthly subscription for? Or is it convenience?

Netflix is not in the content business. They are in the convenience business. Anyone can provide content but nobody delivers convenience the way Netflix does. That’s because their focus is and always has been the user experience. Convenience, the user experience, is why customers pay Netflix. If Netflix had followed the Content is King strategy it would have been Blockbuster.

…Don’t focus on cranking out content. Focus on understanding what the user wants and your content strategy follows.

I cannot overstate the importance of understanding that the user experience underlies many of Google’s important decisions related to its algorithm.”

These are not new ideas that I’m presenting, but they were ahead of their time, maybe still are. Yet, they are as relevant today as they were in 2015 or 2004. Some are saying they are more relevant today because of AI Search.

Featured Image by Shutterstock/Mer_Studio

https://www.searchenginejournal.com/an-easy-digital-pr-strategy-for-ai-seo/581710/

Web strategy in the AI era has a strange shape. You spend your days optimizing for systems nobody will let you look inside. You publish, you watch the traffic move, and when an AI answer surfaces a competitor instead of you, there is no panel that explains why. So when a regulator forces one of these systems to show its work, the question gets concrete: What actually changes for the people who build websites? After reading the order, the honest answer is two things at once. The recourse is real and worth taking seriously. The work it points to is not new.

A Regulator Is Forcing Google To Explain How It Ranks

On June 17, the UK Competition and Markets Authority used Google’s Strategic Market Status designation, granted last October on the basis that Google handles more than 90% of UK search, to impose two binding rules. The first matters to anyone with a website. Google has to rank organic results by “objective and non-discriminatory criteria,” and the regulator wrote that this applies inside AI Overviews, not only the 10 blue links. Google also has to give businesses real transparency into how ranking works, advance notice before major changes to its ranking systems, and a documented process to raise complaints. It has six months. “Step by step, we’re ensuring that Google’s search services work better for businesses and consumers across the UK,” said Will Hayter, the CMA’s executive director for digital markets.

For 25 years, the ranking system was something you inferred from the outside, never something you could question from the inside. Advance notice of changes and a real complaints process is recourse web professionals have never had, and “objective criteria” is a promise that the unexplained demotion has to end. It is UK-only for now, Google will contest it, and nothing is live for six months. But rules like this rarely stay in one country, and the direction is not ambiguous. The layer that decides whether your website is seen might end up being exposed.

Opening The Box Would Change Less Than You Hope

Now run the thought experiment all the way. Say the order goes further than anyone expects, and you could read the exact criteria that decide what gets surfaced and cited, across every engine, not only Google. What would you actually do differently?

You would probably change less than the excitement suggests. Transparency would settle many arguments, sure. It would end the seasonal debate over whether llms.txt does anything (the latest large-scale data says it does not), whether schema markup is a citation cheat code (a controlled study says it is not), whether stuffing a page with “best in class” claims earns the recommendation (it earns the citation and loses the recommendation to the competitors you named). Seeing the rubric would kill the folklore. It would not change the work. A system reading your website still has to find the answer, parse it cleanly, and have some reason to trust it. Whether you can see the criteria or not, the page either presents its substance in a form a machine can extract, or it hides it behind something the machine never runs.

That through-line sits under every one of these stories. A court in Munich ruled in May that Google’s AI Overview is Google’s own speech, which Google can be held liable for. The AI answer is being treated as a product with an owner and rules. None of that touches the one input you fully control, which is whether your content is legible to the thing doing the answering.

Audit What A Machine Can Read On Your Website Today

Waiting for the box to open is the wrong instinct. That is someone else’s six-month timeline, in one country. The right move now is to audit what a machine can already read on your website, and fix what it cannot.

Run three checks, in order.

1. Rendering First: Does your meaningful content exist in the HTML a system receives, or does it depend on client-side JavaScript that most AI fetchers never run? Load your most important page with JavaScript disabled and see what is left.
2. Structure Next: Can the answer to an obvious question be lifted off the page as a clean, self-contained passage, or is it buried in narrative that only resolves for a human reading top to bottom?
3. Verifiability Last: Are the facts that define you, who you are, what you sell, what is true about it, stated plainly and consistently across your website, or does the machine have to take your word for claims it cannot confirm anywhere else?

That is machine-first work, and it is the same work whether Google is forced to publish its criteria or not. It’s upstream of every ruling, which is why it will survive all of them. A website a machine can read, parse, and verify wins in the opaque version of this world and in the transparent one. The only thing transparency would add is proof you were right.

So, when that black box opens is not on your roadmap. A regulator or a court could force that, in their country, on their clock. What should be on your roadmap is whether the answer to a real question about your business sits in your HTML right now, in a form a machine can lift out and trust. You don’t need anyone’s permission to do that.

More Resources:

This post was originally published on No Hacks.

Featured Image: Roman Samborskyi/Shutterstock

https://www.searchenginejournal.com/you-cant-see-how-ai-ranks-you-so-build-what-it-can-read/580048/

This post was sponsored by FirstPromoter. The opinions expressed in this article are the sponsor’s own.

For years, software companies have published pages that rank the best tools in a category and place their own product at the top. The tactic was cheap and easy to scale, and for a long time it helped shape what buyers saw.

In AI search, comparison listicles backfire. Google’s AI Overview quotes the listicle as a source, then recommends a competitor from your own cited list.

Your content only gets the citation. Meanwhile, your competitor gets the recommendation and the click. Your competitor gets the sale.

What makes your cited content recommend competitors?

Lily Ray quantified how often a brand’s own listicle earns the citation but loses the recommendation to a competitor.

In research published in June 2026, she analyzed 100 B2B “best [category] software” queries in Google’s AI Overviews and checked the same queries three times between April and June.

The Results

Across the 80 queries that produced an AI Overview, self-ranked listicles were cited 323 times. In 224 of those cases, Google named a brand’s own page, then recommended a rival ranked inside it.

In other words, when a brand’s own listicle was cited, that brand was left out of the recommendation 69% of the time.

What’s the Difference Between Being Cited & Being Recommended in AI Search?

AI search produces two separate outcomes, and only one of them drives sales.

A citation means the engine named a page as one of the sources behind its answer.

A recommendation means the answer told the reader which product to choose.

The recommendation is what buyers act on.

A citation is easy to mistake for progress, because the brand still appears on the screen.

What an engine cites depends on the content of the page. What it recommends depends on what the rest of the web says about a brand: how many independent sites mention it, link to it, and review it.

Your goal should be to increase recommendations.

Why Does Self-Promotional Content Backfire in AI Search?

Google now treats self-ranked pages differently in its AI answers, Ray found.

The brands that win recommendations are the established names the web already covers.

Recommended brands had far more referring domains, and far more mentions across AI Overviews and ChatGPT, than brands that were cited and passed over.

On-page changes can’t fix this. The gap isn’t on the page; the citation-recommendation gap lies within how often the rest of the web covers the brand.

How to Measure Whether AI Search Recommends Your Brand

You can run this check for any category without special tools. Because citations and recommendations carry different intent, the goal is to separate two figures that usually get combined:

• How often your brand is cited (informational intent).
• How often it is recommended (transactional intent).

Step 1: Build Your Query List

Start with the questions a buyer would type, such as “best project management software,” “Notion alternatives,” or “best [category] software.”

Step 2: Record Citations & Recommendations Separately

Run each one in Google and record two things: the pages Google cites as sources, and the products it recommends in the answer.

Step 3: Repeat Each Query

Run each query more than once, since AI answers shift from session to session.

Step 4: Score Your Share of Voice

Then score the share of recommendations won, rather than the number of citations earned.

Step 5: Extend the Audit Beyond Google

The pattern is documented for Google’s AI Overviews, so begin there. Run the same queries through ChatGPT and Perplexity to map which publishers those engines surface for your category.

Ray’s research shows what the exercise produces. For “best LMS for selling courses,” Google cited Oasis LMS repeatedly, in the body of the answer and in the sidebar. Oasis ranks itself number one in that article. Google recommended Kajabi, Thinkific, LearnWorlds, and Teachable instead, each of them named inside the Oasis piece.

Ray found the same split across categories, from CRM to help desk to SEO software.

Finding 2: Do AI Recommendations Come From Coverage You Don’t Publish? Yes.

Ray’s data shows where AI recommendations originate. Google leans heavily on third-party and user sites, with Reddit, Forbes, and YouTube among the most-cited domains. Content independent from the brand earns a recommendation: reviews, comparisons, and walkthroughs published by someone other than the vendor.

How Do You Get More Independent Brand Mentions That Win AI Recommendations?

You need to increase the number of web pages about your product on domains you don’t control, such as more:

• Reviews.
• Comparisons.
• Walkthroughs.

Each of these should be published by third parties. Not one placement at a time, but as ongoing output.

How Do You Do This Quickly?

Give creators a financial reason to publish. When a creator earns money each time their coverage converts a customer, they keep writing reviews, updating comparisons, and publishing walkthroughs, without you commissioning each piece.

You can start with a handful of creators and a revenue-share agreement. What that produces is coverage. What it does not produce, on its own, is consistency.

How Do You Keep A Consistent Flow Of Mentions?

Run an always-on channel: an affiliate program. Paying creators piece by piece gets you a review here and a comparison there. Mention velocity stays flat because every new URL requires new outreach. Consistent output takes structure: recruiting good partners, tracking what each one produces, rewarding the ones who perform, and paying them on time. An affiliate program is that structure.

Affiliates are third parties who earn a commission when a customer they refer makes a purchase. They include niche site owners, YouTube reviewers, newsletter writers, and media publishers. To earn it, they write reviews, record walkthroughs, and publish side-by-side comparisons on their own sites and channels. That content is what Google draws from when answering a “best [category] software” query.

Proof of Concept: The Brands That Dominate AI Answers Already Run Programs At This Scale.

Run any “best [category] software” query and the same names recur. Behind them sit networks of third-party sites reviewing and comparing those products, earning a commission on the customers they refer. Their referring domain counts keep climbing because the program funds new coverage continuously.

Programs built for editorial output win; programs built for referral volume don’t. A program aimed at raw referral volume tends to attract coupon and deal sites, which drive clicks but rarely publish the editorial content AI Overviews cite. A program aimed at AI recommendations recruits partners who write and review for a living, and favors partners with real audiences over partners who only distribute discount codes.

Telling a strong partner from a weak one takes judgment. The signals worth checking are long-term organic search performance, credible mentions on sites the partner doesn’t control, and a presence across more than one platform. Partners who rank well in AI Overviews usually have that track record already.

“Affiliates are one of the biggest sources of AI citations right now, and yet most brands don’t even think about it. A citation in an AI Overview today doesn’t mean much on its own, because we’re seeing AI-generated sites get cited for a few weeks and then disappear once Google catches up to them. So check the organic history behind it first, look at the fluctuations, scroll through the content. And do that for every partner type, not just websites. A YouTube channel or an influencer can end up in an AI answer too, and they need the same check.” – Tautvydas Vasiliauskas

A referring domain earned this quarter doesn’t keep earning on its own. The brands that hold AI recommendations are the ones whose third-party coverage keeps growing, and that output depends on partners staying active.

Partners stay active when the program is run well. Each task involved is simple. Done by hand, together they consume the hours the program was supposed to save.

Time is not the only thing at stake. AI systems draw recommendations from the pool of referring domains that mention a brand. Low-quality affiliates and self-referrals pollute that pool., and when they do, the citations a program earned stop counting in the brand’s favor.

Keeping the pool clean requires detecting fraud, vetting partners, and blocking self-referrals, continuously, not as a one-time cleanup.

This is the operational work FirstPromoter handles. It tracks each partner’s performance and ties it to revenue, so you can see which partners produce sales, and which produce the coverage AI Overviews cite. It keeps partners motivated with contests, performance tiers that pay higher commissions, one-time placement fees, and target bonuses. Payouts run on a scale from do-it-yourself to fully managed, and setup requires little to no developer resource.

The software won’t choose partners or brief them; that judgment stays with you. It handles the operations, so the coverage keeps compounding without constant hands-on work.

Stop Building Content That Benefits Your Competitors. Start Building Connections That Reinforce Your Brand.

The self-ranked listicle had a good run, and that run is ending. In AI search, Google gives the recommendation to the brands the wider web already trusts, and it builds that trust out of independent content.

An affiliate program is one of the most direct ways to produce content that gets you brand recommendations, and you pay for it based on results rather than adding headcount. It’s worth considering whether you’re starting a program or already run one. FirstPromoter offers a free trial to test the approach.

Image Credits

Featured Image: Image by FirstPromoter. Used with permission.

In-Post Images: Images by FirstPromoter. Used with permission.

https://www.searchenginejournal.com/ai-search-recommending-competitors-firstpromoter-spa/581579/

Last year, 2.5% of Writesonic’s leads came from AI search. As of March, 35% do.

Samanyou Garg, Founder and CEO of Writesonic, showed the system behind that number in his Search Engine Journal webinar: agents surface what moved across every search platform, practitioners prioritize and act. “AI search didn’t necessarily kill SEO, but it turned it into an engineering problem,” he said.

The session covers new citation research and 5 lessons from the field, including the 6-stage loop his team runs on every published page and the workflow that wins citations on pages you don’t own.

Watch the full webinar on demand.

96% Of AI Citations Point To Pages You Don’t Own

Where do AI answers pull citations from? Mostly from pages outside your website.

In Writesonic’s latest research, 96% of AI search citations pointed to third-party sources: Reddit, YouTube, forums, industry publications. A few months ago that figure sat near 80%.

Model updates reshuffle the mix. Reddit, forum, and YouTube citations all jumped between GPT 5.3 and GPT 5.5; the multipliers are in the session.

“You need to make sure you are not putting all of your eggs in one basket, like your own website or a specific website,” Samanyou said.

Action item: Find the pages cited for your target prompts where competitors appear and you don’t. Samanyou demoed an agent that builds that outreach list, author contacts included.

How Long Does An AI Citation Actually Last?

Writesonic measured the lifespan of more than 150,000 citations. The average is shorter than most content calendars assume.

Citations cycle: models rotate in fresh sources, and one model update can hand your spot to a competitor. “It’s a very volatile thing, because models are probabilistic by nature,” Samanyou said.

The exact lifespan number, and the refresh-and-diversify play his team runs when citations rotate out, are in the session. Watch it on demand.

What Goes Into An SEO Agent: 4 Layers

What does a working SEO agent consist of? 4 layers: identity, knowledge, skills, and tools. Samanyou opened real example files for each.

Lesson 1 sits on top: recruit experts, don’t replace them. His team builds expert files, second-brain documents that capture how a named practitioner thinks; their positioning agent runs on April Dunford’s frameworks.

“It’s like a team of junior interns working with you. But those interns are the best ones in the world who have learned from the experts, have access to all the data, understand everything about the domain,” he said. A practitioner approves everything before it ships.

Action item: Build one specialist agent before you build a team of them. The session shows the expert-file method, from deep research to finished markdown.

What Is Closed-Loop SEO? Ship, Verify, Iterate

What is closed-loop SEO? Treat every published page as an experiment: confirm Google indexed it, confirm it ranks or earns citations, feed the result into the next fix.

The live poll showed the gap: most attendees measure nothing, or measure and act on none of it.

Samanyou’s team scores every page with a business impact potential formula: 4 weighted factors that turn a 100-page backlog into a ranked work queue. The 4 factors and their weights are in the session, along with both live dashboard demos.

“Diagnosis is cheap now,” he said. “The main thing is execution.” Watch the webinar on demand.

Q&A: Most Helpful Questions From The Webinar

Q: What should you automate first, and what should you never automate?

Samanyou answered: Connect your existing data sources first and automate the proactive detection loop. Never automate the send: “It should be semi-autonomous until you have a human verifying and testing everything before it goes live.”

Q: On-page or off-page: what actually moves the AI visibility needle?

Samanyou answered: Both, tilted off-page: “I would say maybe 60% focus on off page, 40% on on page.” Rebalance toward on-page once your own pages start earning citations.

Q: How do you create an expert file for an agent?

Samanyou answered: Build a second brain for one named expert: deep research on their published frameworks, or their best talk transcripts synthesized into a single structured file. “It should not be like 10,000 words of text, it should be structured in a proper way, so that the model is able to consume it and use it for any new task.”

Q: How do you know which leads came from AI search?

Samanyou answered: Self-attribution plus verification. The demo form asks “Where did you hear about us?” and sales calls double-check: “There might be maybe 10 to 20% bias where people are just randomly selecting something. But still, it gives us a good indication.”

Watch The Full Webinar

The full session holds the citation lifespan number, the 6-stage loop walkthrough, the business impact potential weights, the citation-gap outreach workflow, the expert-file method, and both live demos.

Watch it on demand.

https://www.searchenginejournal.com/seo-study-5-lessons-from-running-ai-agents-across-every-search/581663/

This post was sponsored by Uberall. The opinions expressed in this article are the sponsor’s own.

Should I add more AI tools to manage local listings and reviews, or is that making it worse?

Who should own AI search visibility across all our locations?

The ideal multi-location marketing world is one where agentic AI fixes duplicate listings, responds to customer reviews, analyzes sentiment, and spots optimization opportunities before the marketer can say “GBP.”

However, what multi-location brand CMOs actually have, in today’s far less ideal world, is layers of disjointed AI and marketing tooling creating an unclean and unclear infrastructure.

This lack of infrastructure makes it nearly impossible to track overall ROI.

An Uberall survey last year revealed that only around 1 in 4 location marketers can show the impact of their location marketing on sales; I’ll bet that with varying levels of AI tool adoption since that survey, this issue hasn’t improved; if anything, it’s been exacerbated by it.

The AI understands what needs prioritizing and resolves it in the background while teams focus on their marketing for multiple locations. It squashes impatience or uncertainty surrounding ROI reporting because its model is built on delivering and visualizing real-time attributable location performance: bookings, table reservations, foot traffic. The clean and clear data that stakeholders wait for.

The results of ill-equipped and layered martech tooling are bleak for local visibility:

• Business listings are managed ad hoc per platform, creating inconsistencies with critical data
• Reviews are left unanswered or sporadically answered, breaking down customer trust and engagement
• Local pages are disconnected from social and inventory systems
• Content is outdated or generic, weakening relevance to local search intent
• Website performance is deprioritized, causing friction for users, search engines, and AI crawlers

Today’s real ideal world is about bringing some sense back to the location marketing stack. It will deliver a combination of that sought-after AI orchestration layer, omnichannel search visibility across locations, and the even more sought-after ROI numbers. It’s the Chief Marketing Orchestrator who will lead it.

Step 1. Decide Who Your Chief Marketing Orchestrator Will Be

Value won’t come from simply plugging data into an LLM. 89% of leaders said their tech investments haven’t fully delivered, with integration complexity the top reason.

Instead, it comes from plugging all your multi-location marketing data into an orchestration layer that implements the nonnegotiable context engineering tasks, making sure every location’s data and signals are structured for any search system customers are using to discover local businesses.

Someone needs to do this, and that person becomes your Chief Marketing Orchestrator (CMO). And, luckily, it’s a new evolution of a Chief Marketing Officer.

The Key Responsibilities of a CMO

The Chief Marketing Orchestrator (CMO) must decide which tasks require human sign-off. Where are the trade-offs? Who owns AI discoverability at a brand and location level? Where can they relieve their team from operational workload and reallocate them to tasks that influence revenue — turning sentiment analysis into actionable reports for operations, or producing content that drives local engagement? It’s not just a technology story but also a leadership story.

Any CMO who is truly passionate about what they do for their multi-location brand doesn’t want to blindly outsource every single task to an AI agent. They want to trust the performance numbers and location marketing initiatives they’re reporting back to stakeholders. And they most likely want to feel in control of compute costs.

At a time when every marketer and every leader is urged to own AI, this often means no one owns the outcome. A streamlined stack with an AI orchestration layer changes that, in that the platform owns the execution and analysis, the CMO owns the overarching strategy, and their team owns the human approvals and guardrails.

This is the principle Uberall’s agentic AI, UB-I, is built on: The marketer remains in control — governing the AI’s output, not just guiding or prompting it.

A CMO investing in the right people to govern agentic AI is a CMO focused on output, not adoption.

Try doing this manually across 50 locations:

1. Open each location’s profile across GBP, Apple, Bing, and relevant directories. Check for formatting inconsistencies, missing attributes, and incorrect hours.
2. Draft a review response for every pending review — starting with the negative ones — matching your brand’s tone and guidelines.
3. Audit each location for missing business descriptions and generate copy that reflects the right local keywords and service context.

That’s the daily baseline. At scale, it’s unsustainable — which is exactly the workload UB-I handles before the team logs in.

UB-I handles the volume and velocity of local operations that no human team can sustainably match at scale, while flagging anything that requires human judgment before acting. On any given day, that means:

1. Drafting AI-generated replies for all pending reviews, according to strict brand guidelines, prioritizing negative reviews first.
2. Correcting name and address formatting to each directory’s requirements, preventing sync failures, and suppressed visibility.
3. Generating missing business descriptions, attributes, and special hours from location data

The team logs in to approve, not to discover what’s broken. Each of these is context engineering in practice — making location data usable for both human and AI-powered search, at a scale no team can manage manually.

As globally recognized innovation strategist Shawn Kanungo puts it: “The companies I am watching win are not the ones optimizing the ROI of existing workflows. They are the ones using agents to do things that were previously impossible at any price.” The efficient orchestration of local marketing tasks across multiple locations has always been impossible at scale — and this orchestration layer is exactly what 99% of senior marketers say would be “valuable” or “very valuable,” according to an Uberall survey.

The true value here in implementing an AI orchestration layer to manage omnichannel presence isn’t to optimize the efficiency of existing local marketing workflows — it’s in enabling what was impossible for marketers to achieve at scale in an eight-hour workday. The workload that 61% of CMOs and VPs at multi-location brands currently describe as “complex” or very “complex” — tracking AI visibility, managing location data and listings, monitoring and responding to reviews, and posting local content on social media.

Step 2. Pivot From Finding New AI To Restoring Search Visibility

As I see it, the solution CMOs will want to implement is to stamp out the ROI-burdening exploratory agentic AI projects and focus on operating with it. Because the prize that comes from operating with it well is attractive for multi-location brands, who need to work quickly to restore declining traffic amid zero-click searches.

Reports indicate that revenue is increasing for brands as customers discover them via AI search — Adobe reports a 254% increase in revenue per visit for the retail segment. It’s no wonder stakeholders are more interested in SEO and GEO performance than ever before.

Let’s imagine a multi-location brand as a building with 200 rooms, each hosting its own party. The furniture hasn’t changed, the walls haven’t changed, the infrastructure hasn’t changed — but there’s a new entrance to the building, one that seems to be a shortcut for guests intentionally looking for you. The other entrances are still in use too. You want to maximize access through every single one so more people find the right room, have a good time, and come back for the next one. You don’t hire someone to manually bring guests to each entrance. You invest in technology to put up signals that do the work for you, so your team can focus on the experience inside the rooms.

Context engineering is what builds those signals. It’s when AI can orchestrate how brands make their digital footprint machine-readable, consistently accurate, technically discoverable across multiple surfaces, contextually relevant, and socially validated — without individuals needing to unpeel layers of tech stack insights.

Implement The 4 Pillars Of Location Performance Optimization (LPO)

If visibility on any search or marketing channel improves, every other location performance pillar improves: engagement, reputation, and conversion. These are the four pillars of Location Performance Optimization (LPO), a revenue-first framework I spoke about at brightonSEO in October 2025. LPO connects a brand’s digital presence to commercial outcomes by activating location data and signals across these performance pillars:

• Visibility: Every location is accurately represented across all relevant discovery surfaces (website, Google, Apple, Yelp, Bing, industry directories).
• Reputation: Trust is reinforced through ratings, regular reviews, and customer resolution.
• Engagement: Local content — posts, photos, offers — signals fresh business activity and relevance for high-intent customers.
• Conversion: Customers can take clear action — bookings, directions, and click-to-calls.

An AI agent that implements these LPO measures to attract more customers, reach new audiences, and influence revenue isn’t exploration. It’s a hard-ROI workflow that pays for the program; they’re the crucial layer that restores and increases search visibility, customer acquisition, retention.

So, when the board asks about AI ROI and local marketing performance, this new CMO doesn’t just demonstrate AI adoption; they justify AI investment to continue to fund their operations. The gap between the brands measuring real ROI and the companies pretending to — or being preoccupied by their complex local marketing stacks is wider than ever.

How To Shift From AI Experiments To ROI-Driven Operations

EY described the moment we’re in well: moving from vibe to value. The “vibe” phase was every company exploring AI — experimenting, piloting, racking up compute costs, layering up their tech stack — and either still being in that phase or having concluded it with the frustration of not knowing how to progress to real, quantifiable returns.

Marketing leaders at multi-location brands, like the Chief Marketing Orchestrator, must adopt and govern agentic-AI-powered stacks that are less exploratory and more ROI-driven. These are stacks that are sensible, streamlined, and enable teams to do things that just weren’t possible before, like logging in to approve fixes, not to discover or prioritize what’s broken. And that approval might not happen before a marketer can say “GBP,” but it’s the orchestration layer — the added AI — senior marketers and leaders are looking for.

Find out how to use Uberall’s UB-I agent for multi-location marketing for your operations

Image Credits

Featured Image: Image by Uberall Brand Studio. Used with permission.

https://www.searchenginejournal.com/local-search-marketing-trends-uberall-spa/581648/

Google is introducing a new property type in Search Console called platform properties, which lets you monitor how social media and video posts perform in Google Search and Discover. This feature supports platforms such as Instagram, TikTok, X, and YouTube and is available even to creators who don’t have their own websites.

Moshe Samet, Product Manager Lead for Search Console, announced the feature in a Search Central blog post. Once an account is connected, a platform property shows which search terms lead people to your posts and how your audience interacts with that content.

The Reports You Get

Each platform property includes the reporting you would expect from a Search Console property, tailored to social and video content.

The Performance report displays total clicks, impressions, and other related metrics, allowing filtering and sorting to identify which posts and queries generate the most traffic. You can also export the data for further analysis in other tools.

The Insights report provides an overview of recent traffic patterns, your most successful posts, and the ways users find your account on Google.

Achievements monitor progress toward milestones, such as surpassing a new total click threshold from Search within a 28-day period.

How To Add a Platform Property

Setting one up involves following Search Console’s verification process: open Search Console, navigate to the verification page or property selector, select Add property, then choose Instagram, TikTok, X, or YouTube and follow the on-screen instructions to authorize the connection.

How This Differs From Search Profiles

Platform properties are distinct from Search profiles, which Google introduced in June as public profile pages for qualified creators and publishers. A Search profile is a shareable page that consolidates a creator’s content for followers. In contrast, a platform property focuses on analytics, showing how those posts perform in Search rather than directly exposing them to an audience. The current feature builds on a December 2025 experiment that initially integrated social-channel data into Search Console.

Why This Matters

You can now track performance across social media and video platforms, alongside your website’s Search performance. This feature allows creators who’ve never had a verified site to see how their posts gain visibility in Search.

Looking Ahead

Platform properties will be gradually available over the next few weeks, so the option might not be visible in your account immediately. Google is initially launching with four supported platforms and directs creators to its help documentation for setup guidance, along with providing a feedback link in Search Console and the Search Central Community.

https://www.searchenginejournal.com/google-search-console-adds-social-video-platform-properties/581634/

A founder pulled up his experimentation dashboard for me last month, proud of it. Forty-one tests running. I asked him to name three that had changed a real decision in the past quarter. He went quiet, scrolled for a while, and landed on one. Maybe.

He isn’t careless. He’s just early to a problem that’s coming for every growth team. The hard part of running an experiment used to be building it. You briefed a designer, waited on ad variants, wired up the tracking, built the page. A week of work to get one test live, with maybe an hour of real thinking behind it. The week of building is gone now. He can launch 40 tests in the time it once took to launch one, so he does, and almost none of them teach him anything.

Volume was never what held teams back. What held them back was telling a real result from random noise, and finding the nerve to kill the losers before they drained a budget. AI solved the cheap problem and left the expensive one sitting exactly where it was. Then it handed everyone a faster way to be wrong.

So, here’s the rule that matters now. The framework you want is the one that gets harder to pass as the tests get easier to run.

What Got Cheaper

The asymmetry I wrote about in team building runs straight through the experimentation pipeline. Spinning up variants costs next to nothing today. Writing a hypothesis worth testing costs what it always did. A model will size your test in seconds and draft the weekly readout in a minute, and it still can’t tell you whether to believe that readout. That takes a person who has been burned by enough pretty curves to distrust the next one.

Point the AI at the production work and keep a clear head on the hypothesis, the design, and the kill call, and the whole thing compounds. Point it at all of it, and you’ve built a machine for shipping noise faster than you can catch it.

Start With Fewer Bets

My first move with a new team is to shrink the test backlog, not feed it. Ask a model for ideas, and it will cheerfully hand you 200. A list of 200 unranked ideas isn’t a strategy. It’s a way to feel busy while the bets that matter wait their turn. The work is choosing the five that count this quarter and saying no to the other 195 out loud, where the team can hear it.

We rank every idea by three questions:

• How big is the win if it lands?
• How sure are we going in?
• What will it cost to run?

Cheap, high-confidence, high-upside ideas go to the front. The one a founder saw on LinkedIn at breakfast waits in line like everything else, unless it clears the same bar. The scoring sheet isn’t the discipline. The discipline is killing a good-sounding idea before it eats three weeks.

One client wanted to tear out his whole onboarding flow on instinct. It scored badly on confidence and worse on cost, so we ran a three-screen test against the flow he already had. His instinct was wrong. The cheap test bought back a quarter of engineering time he was about to set on fire.

A model can write the ideas and even rough out the scores. It cannot tell you which bet your company can afford to get wrong. That call is yours.

Build The Test So The Answer Counts

Most experiments that “fail” never had a chance to succeed, because they weren’t built to answer anything. A clean test moves one variable against a real control, runs to a sample size you fixed before you started, and keeps a guardrail on the number you refuse to harm. Change the headline and the layout and the audience at once, and a lift just shrugs at you. You’ll never know which move did the work. Read the result on day two because the line is climbing, and you’ve promoted noise to strategy.

This is where AI helps, in a narrow and real way. I lean on it to work out how long a test has to run before it can say anything, to simulate the outcome before I spend a dollar, and to catch the obvious confound I miss at six in the morning. The one thing I never let it do is pick the metric. Hand the goal to a model, and it will find you a gorgeous win on a number nobody pays for, while the number that keeps the lights on slides quietly the other way. The human-in-the-loop rule everyone repeats about AI content holds just as hard for test design.

Run The Machine, Not The Judgment

Here’s where the AI more than earns its seat. The build, the variant permutations, the QA, the resizing, the platform formatting, the rough first draft of the readout: give all of it to the tools. Meta Advantage+ and Google Performance Max churn through creative and bids. GrowthBook and Statsig run the statistics and keep your test groups honest. Google Analytics 4 with Mixpanel or Heap holds the event data. A model can turn raw results into plain English, so your analyst spends the hour reading them instead of formatting slides. I laid out the fuller stack elsewhere and won’t repeat it here.

What never leaves a human: the hypothesis, the metric definition, the judgment of whether a result is real, and the call to scale it or bury it. Hand off the labor. Keep the judgment. Most of this framework lives in that one line.

A Cadence You Can Trust

Going fast with no rhythm just gets you to the wreck sooner. We hold one readout a week. Every live test leaves that room with a single verdict: scale, kill, or iterate. There’s no “give it a few more days” unless the test honestly hasn’t reached the sample size we set. And each verdict goes into a log, next to the hypothesis it tested and what we concluded.

That log does the quiet, unglamorous work that keeps the whole system honest. A year in, it’s why a new hire’s excited pitch gets met with “we ran that in March, here’s what happened,” and why a real win from last quarter doesn’t vanish the week after it ships. Running experiments is cheap now. The log is what turns a pile of them into something you actually know.

One Series B client came to us running north of 20 “tests” a month and trusting hardly any of them. We cut it to six properly powered tests, moved the production onto tooling, and put a single weekly scale-or-kill verdict in front of one decision-maker. Inside a quarter, the hit rate on the tests they scaled climbed from a coin toss to roughly two in three, and cost per acquisition fell 24%. They ran a third as many tests and finally trusted the ones they ran.

How The Budget Really Leaks

The same handful of mistakes shows up in nearly every account, and AI speeds up all of them. Teams call a winner on day two because the dashboard refreshes live and the curve looks friendly. They run tests too small to ever reach significance, then read fortunes in the static. They chase a number the model can nudge while the number that matters drifts the wrong way. And the most expensive habit of all: they never kill anything, so the backlog swells, the spend spreads thin, and no single test gets a fair shot.

None of this is new. AI has just put it on a faster clock, which is the whole reason the framework has to keep its shape under speed.

The Takeaway

The teams that win at performance marketing in the AI era aren’t the ones with the most experiments running. They’re the ones who can still believe their own results when the volume climbs. Cheap execution is a real gift. It pays off only if your standards rise as fast as your output does. Make the system harder to pass as it gets easier to run, keep a human on the judgment, and let the machine do the rest. That’s what holds up when the price of one more test falls to almost nothing.

More Resources:

Featured Image: Prostock-studio/Shutterstock

https://www.searchenginejournal.com/performance-marketing-meets-ai-how-to-build-an-experimentation-framework-that-scales/579854/