Author - Rachael Bews

The State of Search 2025: Why Being Human Is More Important Than Ever

TL;DR

Search is no longer just SEO – Generative Engine Optimisation (GEO) is now essential. AI-generated content without human oversight risks being flagged as low quality or spam by Google. Success in 2025 means creating original, human-led, multi-channel content that earns trust.

Key Takeaways:

  • AI-only content gets penalised – Human input and original value are now mandatory.
  • Search is omni-channel – You need visibility across AI Overviews, social, shopping, forums, and more.
  • AI traffic is rising fast – Optimise deep content and homepages for tools like ChatGPT and Perplexity.
  • EEAT is critical – Build trust with expert authors, real experience, and strong brand signals.
  • Social = search – Use SEO tactics to optimise social content and engagement.
  • New success metrics – Track AI-driven traffic and brand visibility, not just clicks.

Keeping It Real

brightonSEO April 2025 came hot on the heels of Search Central Live in Madrid, where John Mueller (Google’s Senior Search Analyst and Search Relations team lead) revealed that automated or AI-generated content may now be hit with Google’s lowest quality score. 

This (and expanded spam definitions – such as scaled content abuse) have followed Google’s Helpful Content Update (in March 2024) which saw the search engine remove a whopping 40% of content from SERPs. The main message? Content needs to have original value for real people to not be considered spam.

This means that our latest and arguably most powerful tactic yet – AI-generated content – could be penalised without a human in the loop. 

In fact, this was exactly what happened to the Content Manager of a very prominent social media platform I met at Brighton who tested programmatic AI content on a large scale and was penalised – so make no mistake, it happens. 

Search marketers the world over – it’s time to sit up and pay attention.

Dropping the ‘E’ - Search Optimisation is the new SEO

It was Alex Hoffman of Passion Digital who suggested we drop the ‘E’ in SEO.

Bold? Maybe. Accurate? Absolutely.

The truth is, the way people are looking for information is now vastly different than even 12 or 24 months ago. Jon Earnshaw from Pi Datametrics, detailed how our search landscape is now made up of:

  • Social media (TikTok, Instagram and others)
    • In 2024, 60% of Gen Z and Millennials preferred social search to traditional search (Alex Hoffman, Passion Digital).
  • Discussion boards and forums (Reddit, Quora and others)
  • AI platforms (ChatGPT, Gemini, Claude, Perplexity and others)
  • Traditional search engines (Google, Bing, Safari and others)

Though they are increasing in popularity, make no mistake – AI platforms remain untrusted by many, are often outdated, can easily hallucinate and are often still reliant on organic search. But the tech powering AI search is changing rapidly.

Choosing the Best AI for the Job

RAG models (which crawl the web – albeit not in realtime – and prefer new, fresh and regularly updated content) were credited by many at brightonSEO with generating the best content (Veronica Hollier, Tresorit). 

 

When it comes to optimising for AI results, it is worth noting that ChatGPT models with web search pull all of their content from Bing – making a strong case for optimisation here (Josh Blyskal, Profound).

The SERPS, They Are A-Changin

You don’t always see what’s right in front of you, and I think this has been the case for me with SERPs. The truth is, they’ve changed a lot since the days of 10 blue links – all because our search behaviour has changed and naturally, Google wants to keep searchers in its product (we’ll likely see this more and more with the move to Google’s AI Mode). 

Ray Saddiq from Rise At Seven, described how:

  • Positions 1-6 are the new 1-10 because of how far down organic positions have been pushed.
  • Third party listicles are often out ranking brand sites for product/service terms.
  • We should be optimising for every real estate opportunity in SERPs: 
    • AIO
    • Paid (from Google Ads)
    • Shopping (from Google product listings in Merchant Centre)
    • Video + short video (from Youtube, Meta + TikTok)
    • Knowledge panel (from structured data / schema markup)
    • Organic social (from Meta, TikTok, X, etc.)
    • Discussion boards / forums (from Reddit + Quora)
    • Organic results (largely from media, publications + listicles)

The goal is to perform well in the entire search ecosystem – not simply to rank highly in traditional search engines.

What’s more, we need to ensure our brand is seen, not just searched – as only 40% of Google searches result in organic traffic – the rest are zero-click searches (Alex Hoffman, Passion Digital). It’s a similar picture when we consider that presence in AI search results drives very limited referral traffic (Baruch Toledano, Similarweb).



The Future of Search Is Omni-Channel

When we really break down SERP anatomy in this way, it is clear that to perform well in traditional search, SEOs need to collaborate with and inform organic social, PPC and PR strategies in a significant way. 

 

We need to look at our Search Optimisation strategy through an omni-channel lens – review all touch points, identify the gaps and build fixes into our roadmaps and workflows. Search data and intelligence need to be underpin activity across organic social, PPC and PR to be effective.

The Future of Search Is Omni-Channel

It might be clear to us already how organic and paid search can work together to tag-team on keyword optimisation and visibility. True also for link building and PR around certain topics and search terms – Listicles and comparative content are the most cited forms of AI content and most popular in SERPs (Josh Blyskal, Profound).

What is perhaps less clear, is how search engine optimisation can, and should, influence social search optimisation. Alex Hoffman of Passion Digital gave us some really practical takeaways to do just that:

  • Use keyword research principles to optimise hashtags, covering branded, industry, trending, event-based, community-focused, and location-specific tags. Apply insights directly to the social content calendar. Practical example:
    • If keyword research reveals high engagement around #SustainableFashion, #LondonEvents, or #MentalHealthAwareness, incorporate these directly into your content strategy.
  • Treat organic social engagement (comments, shares, saves, clicks) as a form of backlinks, incorporating these signals into your KPIs and performance reporting. Practical example:
    • When reporting, highlight metrics like LinkedIn shares or Instagram saves as indicators of content authority and audience trust—similar to how backlinks traditionally signal content value in SEO.
  • Extend SEO beyond website pages to optimise social content for visibility in search results – prioritise conversational headings, FAQs, short-form video, hooks, compelling visuals, statistics, and original data. Practical example:
    • Publish FAQ-style or how-to videos on YouTube or TikTok and ensure they include conversational, keyword-rich titles – this increases the likelihood of appearing in Google’s SERPs as video snippets.
  • Instead of purely keyword-driven intent optimisation, optimise social content to match how users actually engage on each social platform. Practically, this means:
    • Instagram: Optimise for saves, shares, carousel swipes, story replies.
    • LinkedIn: Optimise content to encourage comments, professional discussion, reposts.
    • Twitter/X: Create short, engaging content designed to encourage replies, reposts, and bookmarks.
    • TikTok: Prioritise watch-time, comments, shares, and stitching content.

It’s no longer a question of whether your audience is on social media, it’s how this content is pulling into AIRPs and SERPs (Alex Hoffman, Passion Digital) and it is pulling heavily into LLMs (Lily Ray, Amsive Digital). Optimising social content and engaging on forums (by posting on popular threads in Reddit for instance) can be highly effective for AI search (Josh Blyskal, Profound).

We also need ways to counteract negative press, reviews and perceptions for this reason by using social listening tools. Negative sentiment management is the key we need to provide accurate and up to date information. It’s time to take your brand reputation and visibility into your own hands (Marcus Tober, Semrush). 

How AI Search Is Changing SEO

So when we consider, as SEOs, where solely our sphere of influence lies, we have to turn our attention to:

    • Shopping (in Google Merchant Centre)
    • Knowledge panel (structured data / schema markup)
    • Organic results (traditional SEO with PR collaboration)
    • AI overviews (GEO)
    • AI search (GEO)

When it comes to the rapidly changing areas of AI overviews and AI search, here’s what we know so far:

  • Gartner predicts AI will account for half of all organic traffic by 2028 and ChatGPT monthly users are projected to overtake Google by 2030 (Baruch Toledano, Similarweb).
  • AIOs have reduced clicks by 30%. 99.2% of AIO keywords are informational (as opposed to navigational, commercial or transactional) (Ryan Law, Ahrefs). 
  • AI results are still very volatile, as prompts change and the models learn. Less than a third of top 10 AI search results typically overlap with Google SERPs (Marcus Tober, Semrush). Josh Blyskal at Profound makes this 12%, showing a high degree of variability in the ‘winners’ and ‘losers’ in AI search.
  • 98% of AI traffic comes from three chatbots – ChatGPT (79.5%), Perplexity (15.6%) and Gemini (4.8%) (Ryan Law, Ahrefs).
  • Outgoing traffic from ChatGPT is rapidly increasing (up 1282% YoY 2023-2024), however LLM mentions do not guarantee referral traffic (Baruch Toledano, Similarweb).
  • All 4 major LLMs are more likely to send users to a site’s homepage than Google SERPs, therefore homepages need to be optimised for all user journeys (Baruch Toledano, Similarweb).
  • AI traffic generally prefers articles (including PDFs!) and core website pages (homepage, about page, pricing page, etc.) (Ryan Law, Ahrefs).
  • ChatGPT seems to be pulling more information from pages level 3-4+ (deep insight pages) (Baruch Toledano, Similarweb).
  • ChatGPT prompts are typically much longer than Google searches (Baruch Toledano, Similarweb). Changing search behaviour from information retrieval to question answering in this way reduces the user’s cognitive load and could be arguably why AI platforms are so sticky (Gareth Simpson, Seeker).
  • As a result, Google searches are becoming more navigational. But this is okay, as we know branded searches are also an important EEAT signal for AI and traditional search (Baruch Toledano, Similarweb).
  • Where Google is converting branded search terms into site traffic, ChatGPT is priming searchers – building trust and making them more likely to engage on site. These users are often highly qualified (Lily Ray, Amsive Digital).
  • As of Q4 2024, ChatGPT traffic is much more engaged than Google traffic, spending 3 minutes more on site on average (8 minutes vs 11 minutes) and visiting more pages (11.6 vs 13.3) (Baruch Toledano, Similarweb).
  • The gap between conversions from both types of referral traffic (Google and ChatGPT) is also narrowing. Therefore we want to embrace AI traffic and double down on it (Baruch Toledano, Similarweb).

How Do We Optimise for AI (Aio and Airps)?

Here’s the good news. Traditional SEO (optimising for Google SERPs rankings and top 10 visibility) is still highly relevant. Following SEO best practice can stand you in good stead for good LLM results (Marcus Tober, Semrush).

That said, let’s explore specifically how to optimise for our un(der)tapped areas of potential – AI overviews (AIO) and AI results pages (AIRPs):

You are what you EEAT

EEAT is the holy grail of high quality content for both traditional and AI search. According to Emma-Elizabeth Byrne from Gentoo Media, it requires:

  • Experience – lived-experience in the content subject matter.
  • Expertise – authors who specialise in the subject area.
  • Authoritativeness – credibility of the authors and content (features in other publications, speaking engagements, years of experience, etc).
  • Trustworthiness – is there trust in the website, authors and content from other reliable sources i.e. reviews, testimonials, case studies, etc.

How to Achieve EEAT (Experience, Expertise, Authoritativeness, Trustworthiness)

  

  • Authority, Credibility, Trust: Cite authoritative sources, quotations, and relevant statistics. Acquire high-quality backlinks. Authority and trustworthiness must be externally validated via press coverage, media mentions, client feedback, and backlinks (Jarno Van Driel, Structured Data Consultant).
  • Entity Optimisation: Clearly define key entities (people, places, concepts) using precise terminology and contextual information. Implement robust internal linking to reinforce entity relevance. 
  • Recency Bias: Regularly refactor, update, and republish content to leverage recency bias. Space content publication over time to maintain relevancy (Josh Blyskal, Profound; Ehab Aboud, Uberall). Evergreen content for the win.
  • Direct Answers: Provide concise, direct answers early in your content—especially in meta descriptions, H1 headers, and introductions – i.e. “Top DDOS Platforms 2025” (Josh Blyskal, Profound).
  • Ethical and Transparent Practices: Avoid manipulative practices. Clearly disclose data sources, affiliations, human authors and sponsorships to build user and AI trust.
  • User Reviews and Testimonials: Feature authentic user feedback. High page placement can support conversion (Veronica Holler, Tresorit). Google reviews are shown prominently in AI results (Marcus Tober, Semrush).
  • High-Quality Visuals: Incorporate high-resolution images and interactive features such as 360-degree views where possible. This approach not only engages users but also provides rich data for AI to analyze and present in search summaries (an important and interesting distinction from optimising for organic search).
  • Author Pages and Human Touch: Showcase clearly marked-up author pages featuring real authors. Use “Written by Human, Not AI” badges to reinforce trust (Veronica Holler, Tresorit).
  • Prominence of Google Reviews: Google reviews are highlighted prominently in AI results (Marcus Tober, Semrush).
  • Engage in Peer-Content Platforms: Actively participate in forums and discussion boards like Reddit and Quora to shape brand presence and credibility. Consider activities like AMAs (Marcus Tober, Semrush).
  • Entities and PR Integration: Smaller domains and third-party sources mentioning your brand often have greater visibility in AI results (Marcus Tober, Semrush). Leverage PR to secure these valuable mentions.

User Experience: 

 

  • Content should be engaging, navigable, and user-friendly for users, bots and AI.
  • Optimise headings: Use clear, concise and descriptive headings to organise your content. 
  • Highlight key information: Use bullet points, numbered lists and tables to emphasize important points quickly and easily.
  • Conversation/response: Use a conversation/response style with short paragraphs, bullet lists and simple sentences i.e. FAQs (Colette Masso Del Llano, Contextful).
  • Human in the loop: Ensure content is relevant and correct (especially if AI generated) (Colette Masso Del Llano, Contextful).
  • Build regular check-ins with customer service to tap into feedback loops and ensure UX is optimal (Colette Masso Del Llano, Contextful).

Keywords + Targeting: 

 

  • Expand keyword research: Use semantic analysis to extend traditional SEO research and identify related concepts, addressing user intent more effectively and enhancing content relevance. These can include traditional search terms, long-tail keywords, questions, natural/conversational queries, contextual phrases and related topics.
    • Where SEO matches keywords to content literally, GEO can interpret user intent more accurately, delivering more nuanced and precise responses.
  • Optimise for search intent: in content creation: Aligning with user search intent (commercial, transactional, informational and navigational) enhances visibility on AI platforms.
  • Consider voice/visual search: optimising for these will be essential for maintaining organic visibility and relevance going forward.

Technical optimisation: 

 

  • Optimise load speed (desktop and mobile).
  • Ensure mobile optimisation (AI-friendly readability).
  • Address crawling and indexing issues—ensure crawlability without JavaScript (AI can’t see JavaScript) (Josh Blyskal, Profound).
  • Enhance site security (HTTPS).
  • Structured data/schema markup to improve AI categorisation and visibility in knowledge panels/snippets.
    • Build a 5-star knowledge graph by linking to authoritative sources and showcasing expertise using ‘knowsAbout’ and ‘mentions’ markup (Jarno Van Driel, Structured Data Consultant).
    • Link to high quality sources with a high DA (Jarno Van Driel, Structured Data Consultant).
  • Properly use HTML tags (title tags, meta descriptions, headers).
  • Manage crawling via robots.txt and consider llm.txt (Josh Blyskal, Profound).
  • Use semantic URL slugs descriptive of the content (Josh Blyskal, Profound).
  • Teach the AI about your brand by sharing your content with it. Feeding AI in this way has proven to increase AI traffic (Veronica Holler, Tresorit).

Paid AI Partnerships

 

According to Marcus Tober from Semrush, Perplexity is working with certain partners to promote their businesses. He warns that such partnerships are likely to drive the most clicks from AI search. There is also direct shopping within Perplexity Pro.

At Nu Coton, we fully expect to see paid opportunities in AI search coming down the line.

The Rise of Predictive Analysis

Gareth Simpson from Seeker, described AI search as step 0 in the customer journey – where your brand meets prospects before they even know they need you. 

Anticipating needs before prospects are even aware of them is something that Folashade Uba from A.P. Moller Maersk, delved into when she explored predictive analysis.

By anticipating user behaviour we can plan accordingly, making our tactics more proactive than reactive (i.e. creating content on a topic after it has started trending).

We do this to some degree when we optimise for seasonality within the market. However we can look for other high potential topics in advance, using Ahrefs’ global growth and global forecast functions – allowing us to work on topics 1-2 months before they start trending and share this intel with PR, organic social and paid social teams.

What Does Success Look Like in the New Search Era?

Jon Earnshaw from Pi Datametrics, makes a good point when he asks, how should we measure success in the changing era of search? Is traffic still our main KPI? Is it impressions? Brand mentions? Share of voice? AIRP conversational CTR?

It is still very hard to get this data. There is no GSC equivalent for gen AI engines. Will there ever be? One thing we can do right now is track LLM traffic in GA4 and take credit as an SEO (Lily Ray, Amsive Digital).

 

Josh Blyskal from Profound suggests we start building our muscle now by:

  • Tracking queries
  • Optimising our sites
  • Monitoring and mimicking citations
  • Creating AI conducive EEAT content

The Long and Short of It

Lily Ray from Amsive Digital wrapped up brightonSEO April 2025 very nicely with a clear summary on the current state of search. Here’s what we shouldn’t and shouldn’t do:

Do

  • Appear everywhere.
  • Build a strong, authentic brand across multiple channels.
  • Become a recognised entity in Google’s knowledge graph.
  • Rinse and repeat for all of your authors and experts.
  • Focus on building original, valuable, authentic content with insights and credible opinions for SERPs, AIRPs and AIO.
  • Remember, if tactics seem too good to be true, they usually are.
  • Stay up to speed with Google’s spam policies and be clear on what it deems third party content.
  • Be cautious around parasite SEO practices like pointing to your website from forums (like Reddit, Medium, LinkedIn, Substack, Quora, etc), sponsored guest posts or third party sites in a way that is not clear or transparent.

Do not 

  • Take short cuts.
  • Focus entirely on quick-wins.
  • Use black hat or spam tactics.
  • Fall into the cat and mouse game – you will get caught.

Glossary

  • AI: Artificial Intelligence
  • Gen AI: Generative AI
  • AIO: AI Overview
  • AIRPs: AI Results Pages
  • KEO: Knowledge Engine Optimisation
  • GEO: Generative Engine Optimisation

About the Author, Rachael Bews

A senior performance marketing strategist driving growth for SAAS, tech and cybersecurity brands across the globe. Specialising in search, SEO, GEO and ppc, she has become the go-to partner for senior leaders seeking measurable growth. Rachael is known for her analytical mindset, sharp execution and ability to build performance engines that deliver measurable results.