The New Rules of Search

If you’re looking to grow your business, building brand awareness and driving organic traffic are likely at the top of your list. For years, search engine optimization (SEO) has been the go-to strategy: understand what your audience is searching for, create quality, useful content, publish consistently, and continually optimize your site to climb Google’s rankings.

AI-powered search is rapidly changing that formula. Today, users can get answers from generative AI summaries without ever clicking through to your website. According to a new report from Fractl, AI vs. SEO: How Generative Search Is Reshaping Discovery, Content Strategy, and Consumer Trust in 2025, this shift isn’t a “blip”—it’s transforming how brands earn visibility, trust, and demand. To better understand what this means for founders, we sat down with Nicole Franco, Head of Digital PR and AI Innovation at Fractl, to explore how entrepreneurs can stay competitive.

Let's Dive In.

Your research with Search Engine Land surveyed more than 800 marketers and found that only 4 percent are using AI strategically. What separates surface-level use from real impact?

The majority of marketers today use AI for surface-level tasks, including creating social media copy or drafting emails. While this is good in its own right, it doesn’t really revolutionize the way we do marketing. When we talk about the strategic use of AI, we usually mean using it at every step of a business’s marketing or content creation process. The top 4% of marketers are using AI for tasks such as finding gaps in search visibility, creating content briefs, optimizing for AI citation formats, and monitoring how the brand is used in AI-generated answers.

Great insights! I also found it interesting that, based on your research, organic traffic loss was highest for tech brands. Why do you think that is?

Tech brands are most likely to have content that can be easily summarized or answered by search engines using AI, such as product features or how-to articles. This kind of content is most likely to be answered or summarized in an answer box or a carousel in search results.

CONNECT WITH NICOLE

Makes sense. So what should tech brands do as compared to, say, fashion or lifestyle brands?

Fashion or lifestyle brands are less likely to have that type of content, since it’s much harder to distill into a simple box or answer. For tech brands, therefore, the main strategy should be creating content that is layered with facts but also has opinions, analysis, or exclusive data. In other words, tech brands should be creating content that can be used or quoted in its own right in search results.

AI also prioritizes fresh content. How fresh are we talking?

It’s not necessary for fresh content to be brand new. It’s about having content that’s up to date and relevant to its context. Updated content within 30–90 days is considered fresh for Google’s AI Overviews and other AI-based content generation tools. It is also important to have fresh, relevant content for topics that change frequently, such as technology, health, finance, and regulations. Even evergreen pages benefit from having an audit and small updates that reflect new trends or terminology.

So updating content counts—great. How much of the content needs to be updated for it to be considered fresh?

Updating content is a must for AI to consider it fresh. However, it is also important that the updates are meaningful and intentional. It’s not necessary to simply change a date or a sentence and consider it updated content. To be considered fresh, the content must undergo substantial updates. These include new stats, updated examples, revised sections to reflect current practices, or added sections that answer emerging questions. These updates signal to AI systems that your content is active and maintained. Google also rewards content that reflects intent shifts over time, so it’s about relevance just as much as it is about recency.

What types of content still earn trust and clicks when AI summaries answer the questions first?

While AI summaries do well for simple questions, topics with nuance, emotion, and reality are not their strong suits. This is where content created by people is important. Original research, case studies, industry opinions, thought leadership pieces, and long-form content do well because of their ability to provide context and depth, which AI can’t match. Utility content also works well, such as interactive content, infographics, calculators, and free templates. Content that provides a reason to click beyond just an answer, such as a unique perspective, humor, or a different aesthetic, also tends to do well in terms of higher engagement rates.

Speaking of engagement, reaching new customers is top of mind for many companies. How should B2B and regulated brands reach, say, Gen Z without chasing every new platform?

To reach Gen Z, it’s not about being present on every new social media platform. It’s about speaking their language and showing up in a way that feels authentic to them in the places where they already spend their time. For B2B and regulated brands, it’s about taking complex ideas and making them bite-sized, simple, and easy to understand through visual and interactive content. Depending on the content or product, using platforms like LinkedIn, YouTube Shorts, or TikTok works well for them. Thought leadership content can also be repurposed into Q&A-style content, explainers, or reaction content. Brands that share their stories, their product builds, or their opinions on their values also do well with Gen Z.

For newer companies with smaller marketing teams or leaner budgets, what’s a small change they can make to improve AI visibility?

Add structured data. Implementing schema markup, like FAQ, How-To, and Product schema, can greatly increase the likelihood of your content being included in AI-generated answers. It doesn’t take much programming time but sends a strong signal to the search engines about what your content is about, its structure, and proper citation. Another quick fix is to redesign your existing content into a Q&A or checklist style, mirroring the way users ask questions and the way AI tools scan for answers.

That’s helpful and makes it easy for teams to prioritize. Switching gears a bit: Is there a way brands can reduce the risk of being misquoted or misrepresented by AI?

First, make sure your facts are clear and referenced. Brands should use direct language, define terms, and reference data or links to support their data. Avoid any type of ambiguity that can be taken out of context or misquoted. Secondly, monitor the brand’s presence on AI tools such as Perplexity, ChatGPT, and Google SGE. This is like brand monitoring on social media. If brands find misinformation, they can report it. Lastly, use schema and metadata to reinforce key points and authoritativeness.

Clear, actionable tips—thank you. So how much does social media, or even a blog, factor into the strategy?

They matter more than ever. AI tools are increasingly being built on publicly available content, like blogs, Reddit threads, and social media. A well-organized and frequently updated blog with valuable information can often find its way into a summary or be referenced by an AI. Social signals, like posts on Reddit or Twitter (X), can play a role, too. Think about your blog and social media as part of your social AI footprint.

What still matters: authority, helpfulness, relevance, and structure. What matters less: exact match keywords, keyword density, and over-engineered technical tweaks. AI search rewards content that’s easy to parse and quote, so it’s important to include clear headlines, semantic structure, and direct answers. Schema is more important now than before, along with content that mirrors natural language queries.

What does effective AI quality control look like—keeping in mind that many of our readers have limited resources?

It starts with systems and processes. Build a prompt library to minimize variability. Create a small checklist that all AI-assisted content must pass before publication. It should include accuracy, tone, bias tests, and brand alignment. Finally, designate someone as the last reviewer. Ideally, this should be someone with experience in editing. You don’t need a whole QA team. A well-tightened system can solve most of these problems. Finally, keep all prompts and outputs in an archive. This way, you can easily correct mistakes if needed.

Makes sense. Speaking of resources: Is AI actually reducing workloads or just raising expectations?

Both. Our research shows that 66% of marketers save between 1–6 hours of work each week thanks to AI. However, most of that time is now being reinvested into producing more. So, yes, expectations are rising. AI is making it possible for people to do more, and faster. But unless leaders adjust their expectations, teams risk burning out. It’s extremely important to pair productivity gains with smarter prioritization and clearer boundaries.

In this new world, what SEO rules still apply, and which ones no longer matter in an AI-driven search landscape?

What’s your best tip for smaller marketing teams with limited resources to stay on top of GEO?

Use AI to audit how your brand is represented in AI environments by searching your brand name in Perplexity, ChatGPT, or Google SGE. Pay attention to what pages are surfacing, how you’re described, and if your most important products or messages are being represented. From there, optimize those pages. Unless there’s a major part of your brand that’s missing, there’s no need to create anything new—just update what’s already showing up and make it better.

That prioritization helps. Thanks. People, including me, love to hear about lessons learned. To that end, what are the biggest mistakes you see people making with regards to GEO?

One of the biggest mistakes people make is that they’re trying to optimize for rankings instead of trying to optimize for visibility. The old way of thinking is that if we’re number one, we win. But that’s not really true anymore with AI-powered search. The top link now might be below the fold or entirely ignored. Visibility now means being included in summaries, cited by AI models, or structured in a way that feeds AI systems. Another mistake people make is that they’re thinking of AI-powered search as a technical SEO problem only. It’s a content strategy issue too.

Anything you’d like to add that I didn’t think to ask?

One of the things that I think is really important for people to understand is that the brands that are succeeding right now are the brands that understand SEO is changing in real time. AI isn’t just a tool; it’s a new discovery layer, and the sooner teams begin optimizing for how machines read, retrieve, and summarize content, the more future-proof their visibility becomes.

SEO and GEO Experts Sound Off

Sasha Berson, Co-founder and Chief Growth Executive at Grow Law

“Just a year ago, not one of our clients got leads from AI search. Today, some are seeing 6% of their new business come directly from ChatGPT. That kind of adoption in twelve months is massive. For startup founders, the key is understanding that AI search rewards real authority. You can’t trick it with keyword stuffing or lazy content. It scrapes everything about your business across the internet and decides if you’re one of the best options to recommend. If a startup wants to get found in AI search, the marketing team has to start building real credibility across every corner of the internet. That means publishing expert content under real names, earning reviews consistently, maintaining strong profiles on directories, and making sure their website is technically solid. The way I see it, GEO isn’t some new trick to hack rankings. It’s about proving your business is the best choice, because these AI tools are built to recommend only the few they trust most.”

Stephane Gringer, Co-founder and Lead Guardrail Architect at Voxaic

“LLMs [Large Language Models] work like a panel of experts versus Google SERPs [Search Engine Results Pages], which are a library of links. AI models prioritize information density and vetted authority, making keyword-stuffed 1,500-word blogs a thing of the past. Allowing the LLM to run a citation check by having sources to back up your claims is more effective than the old strategy of amassing as many backlinks as possible. Atomic content is the new content strategy. Bite-sized, straight-to-the-point answers provide LLMs with clear information. Narrative diatribes are for humans and still have their place, but writing for machines is the new semantic structure. Technical SEO is even more important … verbose schema, XML sitemaps, [and] proper headings are critical.”

Back To Content Library