I made a last minute decision to travel to New York City for MozCon 2026. I’m glad I did!

Full disclosure. 50% of the reason i went down was so I could hit my old pizza joint in my hometown in New Jersey. Nothing like a Federici’s pizza followed up with some ice cream from Jersey Freeze.

Now back to MozCon.

The event brought together what seemed like hundreds of search marketers at The Glasshouse for a full day of presentations from some of the industry’s most recognizable voices, including Tom Capper, Amanda Jordan, Debbie Chew, Mike King, Crystal Carter, Chima Mmeje, Beth Nunnington, Paul Norris, Josh Blyskal, and Eric Siu. They covered SEO, local SEO, digital PR, content strategy, AI search, automation, retrieval, and the emerging role of AI agents.

Beyond the presentations, MozCon offered plenty of opportunities to reconnect with people I already knew, meet others I had previously interacted with online, and continue conversations between sessions and during the evening reception.

Now that I’ve had the chance to review my notes, several ideas appeared repeatedly across the day. While many of the speakers approached search from different disciplines, many of their presentations connected through a shared set of challenges.

• Search results are more crowded than ever.
• AI systems are changing how information gets retrieved and presented.
• Automation is creating new opportunities for scale while increasing the need for thoughtful oversight.
• Digital PR, earned media, and third-party validation are becoming more closely connected to visibility.
• Websites are also beginning to serve a new audience: AI agents acting on behalf of humans and other agents.

For this article, I’ll highlight the five themes that helped define MozCon 2026.

Theme #1: Visibility Expands Beyond Rankings

Tom Capper opened MozCon with a presentation titled Billboard SEO: How to Win Google’s Visual Real Estate. He provided a very useful framework for understanding a challenge that has been developing for years: ranking position does not fully describe where a result appears or how much attention it receives.

SEOs typically report rankings through a numbered position. A page ranks first, third, fifth, or tenth. That number provides useful information, but it does not indicate how far down the search results page the listing appears or how much visual space it occupies.

Tom introduced two measurements that help add this context: pixel position and pixel height. Pixel position measures how far down the page a result appears. Pixel height measures how much vertical space the result occupies.

These measurements matter because search results pages contain a heck of a lot more than just ten blue links. Shopping features, local results, videos, images, featured snippets, AI Overviews, and other search features can occupy significant portions of the screen. According to data Tom shared, the top organic result appears below the fold on desktop searches approximately 39%(!) of the time. So, a page can technically rank first while appearing beneath several hundred pixels of other content.

Different industries also produce different search result experiences. Shopping and product features frequently occupy large areas of the page, while standard organic listings tend to be much shorter. Local search results may include map packs, review information, business details, and additional features that move traditional organic results farther down the screen.

This creates an opportunity to evaluate SEO through visual presence as well as numerical position. Rich results can help listings occupy more space. Product features, images, videos, local results, and other search elements can create additional paths to visibility. The goal becomes capturing meaningful real estate across the search experience.

Tom compared this form of visibility to a billboard. A billboard does not require someone to click it to create value. Repeated exposure can contribute to familiarity, recognition, and future consideration. Search impressions can function in a similar way. A person may repeatedly encounter a business through organic listings, local results, videos, AI-generated answers, mentions, or citations before eventually choosing to visit its website or contact it directly.

This idea becomes especially relevant as search journeys become more fragmented. A customer may see a business in Google Search, encounter it again in an AI-generated response, read about it in a third-party publication, and later search for the brand by name. No single interaction receives full credit for the decision, yet each contributes to the outcome.

Rankings remain important. Pixel position, visual size, impressions, mentions, citations, and brand presence add additional context to what visibility now means.

Theme #2: Scale Requires Guardrails and Judgment

Amanda Jordan’s presentation, Local SEO That Actually Works: Some Assembly Required (AI Not Included), examined the relationship between scale and context.

Organizations managing large numbers of locations often rely on templates, automation, standardized processes, and centralized decision-making. These systems make scale possible. They also create the risk of applying the same solution to businesses, locations, and markets with very different needs.

Amanda discussed a familiar pattern in which optimization shifts from outcomes to activities. Teams complete the prescribed tasks, publish the required pages, update the expected fields, and follow the established process. The work gets done, but the business results do not necessarily improve.

Her local SEO case study demonstrated why contextual decisions matter. The work included changes to the primary category, homepage intent, target city, and allocation of content across the website. Together, those changes helped improve Google Business Profile actions.

Each decision required a clear and well-informed understanding of the business and its market. The correct primary category depends on what the business actually does. Homepage intent depends on the services and location that deserve priority. Secondary markets may require internal pages rather than generic location substitutions. Customer language should reflect how people describe their needs rather than how a template expects them to search.

Amanda’s recommendation was to preserve scale while constraining it with guardrails. Systems should include decision points where business context, local knowledge, and professional judgment influence the outcome.

This idea extended beyond local SEO.

Chima Mmeje’s presentation on scaling content emphasized trust, lived experience, external validation, and links to supporting resources. Her PEEL framework encouraged marketers to begin with the audience’s pain, provide evidence grounded in real experience, include external validation, and create useful connections to other resources.

AI can assist with research, ideation, analysis, and iteration. Chima also recommended providing examples and corrective feedback so AI systems have a clearer understanding of the desired output. The process still depends on a human being who can recognize whether the content reflects experience, supports the reader, and deserves trust.

Eric Siu’s presentation focused on operationalizing AI through repeatable skills and workflows. His recommendation to identify time-consuming tasks and convert them into portable skills demonstrated how organizations can capture knowledge and apply it consistently. Multiple skills can then be connected into larger workflows.

The value of these systems depends on the quality of the underlying process. Automating a thoughtful process can improve efficiency. Automating a poorly designed process can reproduce mistakes at a larger scale.

Across these presentations, AI and automation were framed as tools for extending human capability. Effective scale requires clear goals, documented knowledge, contextual decision-making, and guardrails that preserve quality.

Theme #3: Trust Is Built Beyond the Website

Digital PR, third-party validation, and trust appeared throughout MozCon. Debbie Chew’s presentation, 2 Truths and a Lie About Digital PR, began with an important observation: a company’s website is only one source of information about the brand.

Search engines, AI systems, journalists, prospects, and customers encounter businesses across many different sources. Press coverage, podcasts, social discussions, customer reviews, expert commentary, partner websites, industry publications, and community platforms all contribute to the larger picture.

Debbie argued that digital PR has become an important part of making a brand highly citable. For the topics and prompts that matter most to a business, marketers should examine which sources receive citations and which brands appear within them.

This requires looking beyond large, recognizable publications. A niche industry website may have greater relevance and credibility with the intended audience than a broad publication with a much larger readership. The best opportunity is often the publication that reaches the right people and contributes meaningfully to how the brand is understood.

Debbie’s CIDER framework begins with customer intelligence. Brands can use their own data, expertise, and customer insight to identify information others may not have. The next steps include differentiating the idea, executing the campaign, and radiating the resulting asset across additional channels.

A strong research asset can support press outreach, organic social media, paid promotion, newsletters, blog content, partnerships, and sales enablement. The campaign becomes more valuable when it extends beyond earning a single link.

Data stories were a recurring part of this discussion. Original, timely, and relevant information gives journalists a reason to cover a business and gives other sources a reason to cite it. The guiding question becomes: What does this organization know that others do not?

Beth Nunnington and Paul Norris continued this discussion during their panel on answer engines. They defined dominance through citations, mentions, recommendations, alignment with brand claims, and positive statements made by others.

AI systems synthesize information from across the web. Owned channels establish the brand’s position, while third-party sources help validate that position. Consistent messaging across content, media coverage, podcasts, forums, and other channels gives these systems more evidence to work with.

Recency also plays a role. Current coverage and recent discussions may provide stronger evidence of what a business represents today than information published several years earlier. This gives brands another reason to maintain an active presence across relevant channels rather than treating digital PR as an occasional campaign.

Beth summarized this particularly well during the panel:

“You can buy attention, but you can’t buy trust.”

Advertising can place a brand in front of an audience. Trust develops through experience, consistency, expertise, customer feedback, and validation from independent sources.

Chima Mmeje made a similar point when she described trust as content’s greatest currency. Her emphasis on lived experience and external validation connected content quality directly to credibility.

Together, these presentations reinforced a principle that has become increasingly important across both SEO and AI visibility: the information a business publishes about itself represents only part of its digital footprint. Reviews, coverage, citations, commentary, partnerships, and customer experiences help determine whether those claims can be affirmed independently and deserve confidence.

Theme #4: Retrieval Is Becoming a Visibility Layer

Query fan-out appeared once again at MozCon. Mike King described it as one of the most important areas marketers can focus on as AI-powered search develops. His presentation emphasized the need to build content ecosystems that support the broader retrieval processes used by modern AI systems.

A traditional view of search begins with a query. The search engine matches that query to relevant documents, ranks those documents, and presents them to the user.

AI search can involve a more expansive process. A system may interpret the original request, generate related or supporting queries, retrieve information from multiple sources, and synthesize the results into a response. This process is often described as query fan-out.

The original query remains important, but it may represent only one part of the retrieval process. A user asking for a product recommendation may trigger searches related to features, price, comparisons, reviews, use cases, reliability, alternatives, and current availability. A local service query may lead to retrieval related to location, reputation, service coverage, expertise, reviews, and customer experiences.

Each supporting query creates another opportunity for a source to enter the final response.

This helps explain why content ecosystems matter. A single page may provide a strong answer to the primary query while leaving many supporting questions unanswered. A connected collection of service pages, product information, customer stories, research, comparisons, expert commentary, and third-party evidence gives retrieval systems more paths to relevant information.

Josh Blyskal’s presentation on the state of AI search provided examples of how retrieval behavior varies across platforms. He noted that one in four click visits from ChatGPT lands on a homepage, illustrating the importance of a clear and informative primary brand page.

He also discussed Claude’s reliance on Brave Search and the relationship between Claude citations and high-ranking Brave results. His analysis suggested that Claude frequently appends a year to fan-out queries, creating potential value for clearly dated and current content when recency matters.

Google AI Mode introduces its own retrieval and presentation behaviors. Josh discussed how AI Mode may examine and rewrite headings, adding another layer between the content a publisher creates and the way it appears in an AI-generated response.

These examples illustrate why AI visibility cannot be reduced to a single ranking or platform. ChatGPT, Claude, Google AI Mode, Gemini, and other systems may rely on different sources, retrieval processes, ranking systems, and interfaces.

Retrieval has become another layer of visibility. A brand must first become available to the system, appear within relevant retrieval paths, and provide information the system can use. Citations, mentions, and recommendations occur later in that process.

Mike also connected these developments to the growth of agents. As agents increasingly digest web content and complete tasks, content will need to support systems that retrieve, interpret, and act on information rather than simply display it to a human reader.

Theme #5: Websites Are Becoming Interfaces for Agents

Crystal Carter closed the day with How to Get Your Website Ready for AI Agents. Her presentation focused on a growing change in how websites may be used. AI agents can complete tasks on behalf of humans and, in some cases, interact with other agents. That creates a new type of website visitor with different requirements from a traditional human user.
A human visitor can often work around ambiguity. They can interpret visual layouts, infer the purpose of a button, search for missing information, or adjust when a process is not immediately clear. Agents depend more heavily on explicit structure, instructions, accessible interfaces, and clearly defined actions.

Crystal emphasized the importance of making websites easier for agents to understand and use. Clear calls to action, accessible navigation, well-labeled forms, structured documentation, and searchable knowledge resources can all help agents interpret what a site offers and how to interact with it.

Accessibility plays an important role in this process. Crystal noted that agentic browsers may use ARIA tags and other accessibility information to interpret page elements. Many of the practices that make a site easier for agents to use also improve the experience for human visitors, particularly those relying on screen readers or other assistive technologies.

One of the most useful distinctions in her presentation was between declarative optimization and imperative optimization.

Declarative optimization is about telling the agent what it needs to know. This can include relevant information, instructions, standard operating procedures, negative constraints, and reference documents. The goal is to provide enough context for the agent to understand the task, the available information, and the boundaries within which it should operate.

Imperative optimization focuses on enabling the agent to act. This can include allowing it to carry out actions, access APIs, directly access site data, use agentic tools within the website, or connect through technologies such as Model Context Protocol, or MCP.

Understanding this difference is important because comprehending a task and being able to complete it are two different things. A website may provide clear information about how to schedule an appointment, place an order, request a quote, or complete another task. That gives an agent the context it needs to understand the process. An agent-ready experience goes a step further by providing the technical pathways that allow the agent to actually complete the action on the user’s behalf.

This changes how we may need to think about website optimization. Content still needs to explain what a business does, who it serves, and how its products or services work. At the same time, agents increasingly need clear instructions, defined constraints, accessible interfaces, and reliable ways to interact with site functionality.

Crystal also discussed machine-readable guidance such as iims.txt, which may help AI systems identify important information and better understand how a site is structured. Tools like this fit into the broader goal of making information easier for machines to find and interpret.

Mike King’s presentation approached the same development from a broader perspective. He argued that agents will increasingly become systems that consume web content and complete tasks, creating what he described as the machine media era.

That has implications beyond SEO alone. Website strategy, content, technical SEO, accessibility, product design, data, digital PR, and brand all contribute to the environment an agent encounters.

A site may have strong content but poorly structured forms. It may be technically accessible while offering unclear instructions. It may provide useful documentation that is outdated or disconnected from the actions users need to complete. Any of these gaps can limit how effectively an agent understands or interacts with the site.

Preparing for agents therefore requires thinking beyond whether content can simply be crawled or retrieved. Websites increasingly need to support the full interaction: providing the information and constraints agents need to make accurate decisions while also creating the technical pathways that allow them to take action.

That may become an increasingly important distinction as AI systems move from answering questions to completing tasks.

The Expanding Definition of Search Visibility

The five themes that defined MozCon 2026 connect through a broader change in how visibility works. Tom Capper demonstrated that ranking position alone does not describe where a result appears or how much visual attention it receives. Amanda Jordan showed how context and judgment determine whether scalable SEO systems produce meaningful business outcomes. Debbie Chew, Beth Nunnington, Paul Norris, and Chima Mmeje connected trust to earned media, customer insight, external validation, and expertise. Mike King and Josh Blyskal examined the retrieval processes influencing AI-generated answers. Crystal Carter explored how agents are becoming active users of websites.

Together, these presentations describe a search ecosystem with more participants, more interfaces, and more paths between discovery and action.

A customer may encounter a brand in a traditional organic result, local pack, shopping feature, video, press article, podcast, AI-generated response, or recommendation. An AI system may retrieve information from the website, a third-party publication, customer reviews, a forum, or a search index. An agent may then visit the website, interpret its documentation, complete a form, or book a service.

Each interaction contributes to visibility and each depends on a different collection of signals. Rankings remain part of this environment. They continue to influence discovery, traffic, and retrieval. Their meaning becomes more useful when combined with pixel position, impressions, citations, mentions, trust, brand presence, business actions, and the ability of machines to access and use information.

Final Thoughts

As I reviewed my notes from MozCon 2026, the conference offered a practical view of how marketers are adapting to a search landscape shaped by traditional search engines, AI assistants, answer engines, and emerging agents. The presentations did not revolve around a single tactic. They connected around a larger set of responsibilities.

SEOs need to understand how brands occupy visual space across search results. They need scalable systems that preserve context and judgment. They need evidence and third-party validation that support trust. They need content ecosystems capable of participating in broader retrieval processes. They also need websites that humans and machines can navigate, understand, and use.

SEO continues to expand across disciplines that organizations often manage separately. Content, brand, PR, reputation, accessibility, technical infrastructure, data, and user experience all contribute to whether a business becomes visible, trusted, retrieved, and selected. That may be the clearest takeaway I brought home from MozCon.

Search visibility now extends across the full digital ecosystem. Success depends on understanding how all of its parts work together.