Have you ever been asked in an internal meeting, “What’s the actual difference between LLMO and SEO?” and found yourself unable to answer right away?

At Google I/O in May 2026, Google announced its policy of rebuilding search around AI. The protagonist of search results is shifting from a list of links to AI-generated answers, and this shift has suddenly made the question feel much more pressing.

SEO and LLMO share a great deal of common ground, but the changes taking place fall along three axes: goals, the focus of initiatives, and KPIs. This article organizes how the assets built up through SEO can still be put to use, and where new layers need to be added. Note that there isn’t yet an established “correct answer” for LLMO — please treat the framework here as just one current perspective.

What you’ll get from this article

• Be able to explain the differences between SEO and LLMO along the three axes of goals, focus, and KPIs
• Understand which parts of your existing SEO assets remain valuable in the age of AI search, and what needs to be added
• Grasp the key points of how to “place” content so your company becomes the one being cited
• Learn a new way of thinking about measuring effectiveness that replaces visit counts, plus a first step you can take today

The Three Axes: Goals, Focus, and KPIs

① The goal shifts from “being clicked” to “being cited”

The very premise of getting people to visit your site is being shaken. The traditional goal of SEO was to rank highly, draw traffic to the site, and convert that traffic. However, a Semrush study found that 93% of searches in Google’s AI Mode end without any transition to an external site, meaning the old structure is becoming harder to sustain after this overhaul of the search experience. If no one visits, nothing downstream can even begin.

What has emerged in its place is a new goal: being cited by name within the AI’s answer. A Seer Interactive study reported that sites cited within AI Overviews tended to have a 35% higher organic click-through rate and a 91% higher paid click-through rate compared to sites that weren’t cited. The researchers note that causation can’t be confirmed, but it does seem clear that whether a company gets cited is starting to determine its visibility in search.

② The focus of initiatives shifts from “keywords” to “facts unique to your company”

To become the one being cited, what needs to change is where your initiatives put their focus. Traditionally, content and site structure were designed starting from keywords — writing articles to match queries and organizing internal links and titles accordingly. This work is still considered effective; many observers note that foundational elements like technical SEO and E-E-A-T also often serve as judgment criteria for AI.

What’s newly added on top of this is the idea of presenting facts that only your company holds, in a form AI can pick up: pricing, specs, customer case studies, proprietary research data. For example, this might mean a pricing-plan comparison table for a B2B SaaS company, structured spec data for a manufacturer, or a proprietary research report for a consulting firm. An AirOps study found that including three tables on a comparison page tended to raise citation rates by 25.7%, while including eight list sections on a verification page raised them by as much as 26.9%. In other words, beyond the content itself, how it’s presented becomes a key factor.

That said, this isn’t something you finish just by adding a tool. Even with schema.org markup in place, if the underlying information is disorganized or inconsistent across channels, it becomes harder for AI to view it as trustworthy. Organizing information is directly tied to site structure, content operations, and internal information management — so this isn’t just a web initiative, it’s an effort to align information across the whole organization.

③ KPIs shift from “visit counts” to “citations and brand mentions”

The last thing that changes is how you measure results. Under traditional SEO, success was basically measured by visit counts, page views, and conversions. But as more cases arise where the AI’s answer fully satisfies the user and no one ever visits the site, visit counts alone can no longer capture what’s happening.

Candidates that have emerged include the number of times a company is cited by AI (appearing as a source) and how often the brand is mentioned within AI answers. Branded search volume still matters too, and that can be tracked with existing tools like Search Console. Citations and mentions, however, aren’t fully captured by conventional analytics, so a separate measurement approach will likely be needed. That doesn’t mean rushing to adopt a dedicated tool, though — as a first step, you can pick a few key queries, ask the AI those queries yourself, and manually log whether your company gets cited or mentioned. Even this kind of simple, repeated observation will reveal where you currently stand and which way things are moving.

Redesigning KPIs also involves building consensus within the organization. A team that has long used “how session counts changed year over year” as its common language will need real time and effort to switch to new metrics. Having stakeholders agree to run old and new metrics in parallel for a while will reduce friction during the transition. It’s worth starting the metrics discussion alongside tool selection, rather than after.

Summary

The assets built up through SEO are being repositioned along the three axes of goals, focus, and KPIs. The foundation still matters, but a new layer is being added that the foundation alone can’t reach — that’s the understanding that feels closest to reality.

What matters here is that simply adding a tool tends not to be very effective on its own: if the underlying information and operations aren’t in order, a tool alone won’t easily move the needle. That’s exactly why information design, operational workflows, and internal agreement on metrics all need to be aligned together. And where all of this ultimately takes shape is at your customer touchpoints, starting with your website — which page to place which citable fact on, and how to implement structured data. At its core, it comes down to the process of designing and building out content.

We provide end-to-end support

Neuromagic helps companies transform their knowledge and track record into “information that can be found” and “information that gets cited.” We support everything from organizing information so it’s easily cited by AI and designing site structure, through to implementation and checking results after launch.

To get the full picture, start with our downloadable resource (available via the button below): “Toward an Era Where ‘Just Building’ Isn’t Enough to Be Chosen: A Website Strategy Proposal for the Age of AI Search.” Please note that the resource is currently available in Japanese only. If you’re curious how AI-ready your own site is, a free diagnostic is also available.