Recently, we looked at the debate around whether ‘SEO’ needs a rebrand to reflect the increasing emphasis on AI-generated answers that is shifting how users approach search and, as a result, how marketers approach optimisation.
In this follow-up, we’re going to dig deeper into the question of how to optimise for the changes in search behaviour that have emerged as a result of the new paradigm, and how marketers can gauge the success of their efforts.
Moreover, as marketers increasingly think about how to appeal to large language models and even AI agents, how can we be sure not to lose touch with what makes good content for humans? I turned again to our experts in search, visibility, and discovery to get their take.
With search engines and large language models providing more detailed answers in response to an initial query, Charlie Marchant, CEO at Exposure Ninja, notes that “We’re … seeing user journeys speed up.
“People [might] start in ChatGPT to research or get recommendations, then go back to Google to make a final, branded search. AI answers are influencing what they then Google later, which changes how businesses need to think about visibility at every stage of the funnel,” she says.
Irina Ginghină, SEO Director at Modern Citizens, agrees: “AI is designed to guide users through complex decisions faster. So, our content needs to anticipate those deeper, multi-step queries.”
She advises that, “Instead of just providing a single answer, we should offer comprehensive, interconnected information that addresses follow-up questions and different angles of a topic.
“This helps AI build a richer understanding and, crucially, provides more complete value to a human user who is on a longer research path.”
“What has changed is how we search or prompt these generative engines such as ChatGPT,” says Sean Begg Flint, Founder at Position Digital. “Research shows that our prompts are much longer, more detailed and more nuanced than on traditional search – we’re also conducting more follow-up searches. When searching for a product or solution, we may include more detailed information about ourselves and the specific problem we are facing.
“This means that brand positioning and persona targeting is more important than ever before when it comes to content optimisation.”
Begg Flint notes that carrying out ‘AI brand audits’ can give marketers an insight into how consumers find their business and where they appear in answers. “[This] involves identifying the specific prompts your target audience may be searching to find your product or service, analysing the AI-generated answers on all the relevant platforms, and understanding all the sources across the web that have influenced the answers, whether onsite or offsite.
“The goal is to identify any inconsistencies or potential ‘hallucinations’ and then implement a content strategy to align all the content to the target audience. It’s marketing 101 but as SEOs we may not have focused on this as much in the past.”
Ginghină identifies three major shifts on the horizon for marketers as they adapt to the evolving search landscape. “One: Websites as branded experiences.
“AI can surface information, but it can’t replicate the experience of a brand. The role of the website is shifting toward being the place where users look for validation, context, interactive tools, or personalised journeys. Essentially, the things an AI overview can’t reproduce.
“Two, distribution across multiple discovery channels: Relying on Google alone is no longer enough. People discover content on TikTok, YouTube, Reddit, marketplaces like Amazon, and of course through LLM‑powered assistants.
“Social search and short‑form video are making just as big a dent in behaviour as AI overviews. Optimisation now means ensuring [a] brand has presence wherever discovery happens, not just in search engines.
“Three: Because AI shortens the research phase, users who land on our site often arrive closer to taking action. … SEO success isn’t only about traffic anymore; it also includes visibility in summaries and answers, trustworthiness of recommendations (informational, product, or service), and brand presence even in non‑click environments.
“So while the fundamentals of SEO remain constant, the battleground has shifted. It’s no longer just about ranking pages; it’s about ensuring [a] brand shows up, earns trust, and delivers experiences that AI cannot.”
One of the earliest signs that AI answers were impacting the landscape of discovery was a steady drop in traffic and click-throughs from search.
Many readers will have seen graphs that depict what is often called the ‘crocodile jaw’ (or ‘alligator jaw’ depending on your preferred species): a upward, or roughly level, trend in impressions from Google for brands and publishers with click-through falling away below it. And while referral traffic from LLMs like ChatGPT is increasing, it still represents a tiny fraction of overall traffic.
For marketers relying on clicks and site visits to understand the impact of their SEO efforts, this felt like a blow, but how much do measurement efforts really need to be reimagined with the introduction of AI answers into search?
Will Findlater, Group Head of Content at Nelson Bostock, puts forward that measures of success will depend on the objectives of the optimisation programme. “When it comes to brand health, citation volumes in answers to common queries and the sentiment of those citations are important ways to measure the success of GEO programmes.
“But if you’re looking to generate leads, click-throughs and conversions should be measured as well. The good news is that traffic driven by LLMs is of a very high quality compared to traffic from traditional search and tends to come from people who are further along their buying journey. So the volume may be lower, but the intent is often stronger.
“To give a complete picture,” he advises, “marketers should complement channel-specific tracking with broader measures that quantify how their content and PR initiatives are helping to drive brand uplift and organic demand generation.”
For brands and businesses not focused on using organic search to drive sales and lead generation, Exposure Ninja’s Marchant recommends tracking brand visibility and brand sentiment as key markers of success. “Brand visibility … is the new version of ranking on page one,” she says.
“Are you showing up in the answers to prompts you care about being seen in? Are you being recommended in ‘best of’ lists or as a solution to specific problems? Is your website content being cited as a source? Are you being named directly in AI-generated answers? And what position are you showing up in those AI answers?
“The second part is brand sentiment. It’s not just about whether you are mentioned, but how you are being positioned. Are the AI platforms speaking about your brand positively? Or are you being mentioned with negative or neutral language? Is the information they share about your business correct and up to date? Positive sentiment influences trust and plays a huge role in decision-making, especially in early buying stages.
“The brands that succeed in AI Search are positively positioned and consistently referenced or mentioned.”
“Classic traffic metrics still matter, but they don’t tell the whole story anymore,” says Irina Ginghină. She breaks down the measures of success into three areas:
1. Presence
“Before anything else, we need to be visible where discovery happens. That means appearing in AI summaries and answer engines, as well as in social search feeds and traditional results. Without that initial presence, the rest of the funnel never starts.
2. Performance
“Once we’re present, the question is whether those interactions turn into meaningful outcomes. Click‑through rates may be lower, but the visitors who come through are usually further along their journey, having already been filtered through AI.
“Crucially, conversions may now happen in two places: directly on our website or within the platforms themselves (TikTok Shop, Amazon, Instagram Checkout, Google Shopping, etc.). Success here is about ensuring that, wherever the transaction happens, we have created the clarity and trust needed to move a user from intent to action.
3. Equity
“Finally, there’s brand equity. Even if a user doesn’t immediately click through or buy, being cited or recommended by an AI system builds recognition and trust. Each time our brand is surfaced as a reliable answer, it reinforces authority. It’s not always easy to measure, but it influences long‑term demand in ways traffic dashboards can’t capture.
“Taken together, presence, performance, and equity give us a more realistic picture of success,” she concludes. “The fundamentals of SEO remain, but what counts as ‘winning’ has expanded. It’s no longer just about traffic, but about ensuring we show up in the right places, enable conversions wherever they happen, and build long‑term brand trust in the process.”
From the earliest days of keyword-stuffing to rank more favourably in search, marketers have needed to walk a line between appealing to search algorithms and web crawlers and thinking about the best way to serve the people their content is ultimately created for.
The debate about creating human-centric content is therefore nothing new but is as important a consideration as ever in the age of LLMs – particularly when the discussion turns to ranking for things like fan-out queries, which are conducted by an AI tool and not a human.
“Quality and trust signal definitions have become much more sophisticated, but there will always be those who try to game the system, particularly in these early days,” observes Findlater.
“The good will out, I hope – that’s why it’s great that LLMs cite impartial, trustworthy sources more frequently than brand sites, and why broader Discovery Optimisation programmes are going to be critical to keeping … useful information front and centre.”
Judith Lewis, Founder at Decabbit Consultancy, adds that while optimising for an LLM doesn’t inherently need to mean creating less worthwhile content, “I think the use of LLMs to write content may limit the inherent usefulness to humans as it becomes just another regurgitation of content already available.
“Future success will come from content that has been written for people, by people, with a specific goal in mind – such as a review, explanation, etc.
“I’d love to see authorship come back with links to EEAT (Experience, Expertise, Authority, Trustworthiness) proof for authors.”
“The great thing about AI-friendly content is that it’s also incredibly human-friendly,” Marchant points out.
“Optimising for LLMs means your content has to be easy to parse with strong headings, logical flow, and clearly defined answers to natural language questions. That structure doesn’t just help AI understand your content, it also helps readers scan and find what they’re looking for faster.
“LLMs also reward credibility. They prioritise content with first-hand expertise, citations, and transparency about the author. That’s also what builds trust with real people who want to know they’re getting advice from someone who actually knows their stuff.
“And because LLMs use a fan-out process to understand the broader context of a query, your content has to go beyond surface-level answers. It needs to address related questions, add practical insights, and deliver a more complete picture. That kind of depth builds authority and keeps your human readers engaged.”
In sum, “If you’re optimising for AI, you’re also creating content that’s more useful, more trustworthy, and more valuable for actual humans. The two are not in competition. If anything, they’re now inseparable.”
Ginghină offers a reminder that “AI is a proxy, not the end user. LLMs and AI agents are trained to act on our behalf, but they judge content differently to humans: they parse entities, structure, and credibility signals, whereas we look for clarity, trust, and experience. Striking the balance between the two is where real value lies.”
Moreover, she adds that, “AI is highly effective at recycling what’s already out there. The real human value comes when content goes beyond the obvious: offering new insights, original data, or lived experience.
“That’s what lends weight to AI recommendations and ensures that humans using them feel they’re seeing something credible and distinctive.
“…[E]nsuring AI optimisation serves people comes down to designing for both audiences at once: clarity, structure, and entity strength for machines; depth, credibility, and human signals for users. AI may filter discovery, but trusted human voices remain at the centre.”