How to Strengthen and Optimize Your SEO Strategy?

TL;DR

In late 2023 we published an SEO strategy post that treated SEO the way it had worked for a decade: keywords, backlinks, technical health, mobile, voice search. The past two years made most of that framing obsolete. AI Overviews now appear on close to half of all Google queries and are cutting click through rates on affected top rankings by 30 to 50 percent. ChatGPT, Perplexity, Claude, and Google’s AI Mode have become primary research tools for a majority of B2B buyers. Traditional SEO still matters. But SEO in 2026 requires optimizing for two surfaces at once: the traditional ranked list, and the AI generated answer that increasingly sits above it. This post is what we tell clients today when they ask why organic traffic is dropping despite better rankings.

Key Takeaways

• AI Overviews now appear on roughly 48 percent of Google queries. A number one organic ranking earns dramatically fewer clicks than it did in 2023.
• The clicks that do come through convert at much higher rates. Fewer visitors, more intent. That changes how you measure and report on SEO.
• Around 80 percent of B2B tech buyers now use generative AI at least as much as traditional search. Getting cited inside ChatGPT, Perplexity, or an AI Overview is now a real acquisition channel, not a bonus.
• The most under invested area in almost every SEO program we audit is structured content built for AI extraction: FAQ schema, comparison tables, direct answers, and entity clarity.
•  LLM SEO, AEO, and GEO all refer to roughly the same discipline. Ignore the acronym wars. The real question is whether your content is extractable by an AI.
• Most 2023 tactics are not wrong. They are insufficient. Backlinks still matter. Technical health still matters. Keyword targeting alone is no longer a strategy.
• If organic is a serious channel for you, run a monthly citation audit across ChatGPT, Perplexity, Gemini, and Google AI Mode. Most brands have never checked whether they show up.

Introduction

When we wrote the original version of this post in December 2023, we described SEO as if its fundamentals had stayed roughly the same for a decade. That was defensible then. Google Hummingbird, mobile first indexing, voice search: all real, all covered. Looking back at it two years later, it reads like a snapshot of an era ending. The past 24 months have compressed more change into search than the previous ten combined.

In the marketing engagements we have run since, we have watched client programs that ranked number one for their target terms in 2023 lose 40 to 60 percent of their organic traffic in 2025 while holding those rankings. We have watched other clients pick up material inbound pipeline from ChatGPT and Perplexity citations without ranking on Google at all. And we have watched a growing number of RevOps directors ask the same question: our SEO looks fine on paper, so why is pipeline from organic down?

This post replaces our 2023 SEO strategy guide with what we tell those clients today. It is not exhaustive. It is what we prioritize when we redesign an organic acquisition program for the search environment as it actually exists in 2026: AI Overviews, LLM citations, compressed CTRs, and a buyer who researches with an AI assistant before they touch Google.

FAST FACT: Ahrefs data shows Google AI Overviews now appear on 48 percent of all search queries, up from 34.5 percent in December 2025. For informational and how to queries the trigger rate is above 70 percent. Source: DigitalApplied AI Overviews Analysis, March 2026. 

What Changed Since (And Why the Old Playbook Is Dying)

Three shifts have hit SEO hard between our original post and today.

AI Overviews took over the top of the SERP. What sits above the traditional organic list is now often not a link but an AI generated summary. Being cited inside that summary matters more than ranking position three below it. Per Stackmatix’s aggregate of Pew, Ahrefs, and Amsive studies, CTR at position one has dropped 34.5 percent on affected queries, with the worst hit categories (B2B tech) seeing declines above 70 percent.

Clicks compressed. Intent concentrated. The bad news is fewer clicks. The good news, and something we see repeatedly in client analytics, is that the clicks that do come through convert at meaningfully higher rates. Some sources cite 23x lift. In our clients we see something more modest but still large: two to four times higher intent on AI Overview referred visitors. Google is filtering low intent visitors before they reach you.

A parallel search layer emerged. ChatGPT, Perplexity, Claude, and Google’s AI Mode have become primary tools for B2B research. Around 80 percent of B2B tech buyers now use generative AI at least as much as traditional search for vendor evaluation. Being cited in an AI answer is a discovery event that traditional SEO tools do not measure. Most clients we work with have never audited their visibility in this layer.

For teams whose organic strategy still runs on 2023 assumptions, all three of these shifts are visible in the traffic reports. The confusing part is that the reports often look worse than the pipeline actually is. High intent conversions are up while total sessions are down.

How Google Search Actually Works in 2026

The mental model of Google as ten blue links ranked by authority is now inaccurate for most commercial queries.

Here is how a typical B2B query resolves today. The user types a full sentence, often a paragraph. Google returns an AI Overview at the top for around half of these queries. The Overview cites two to five sources, usually pages that already rank in the top ten organically. Google’s own behavior confirms this: the AI Overview pulls citations mostly from pages Google already trusts.

Below the Overview sits a mix of traditional organic results, ads, video panels, and the People Also Ask box, which now often expands to answer the follow up questions the AI Overview did not resolve. The old top three positions still get clicks, but many fewer. Positions four through ten essentially do not exist for informational queries anymore.

What ranks in this environment is content that is structured for extraction. Pages that answer the exact question in the first paragraph, use question format H2 headings, contain FAQ schema, and include comparison tables where relevant. Google is not searching for keyword matches. It is searching for authoritative, machine parseable answers. For the operating model behind how we rebuild organic programs for this environment, see our Digital Marketing practice.

What still matters and always will: backlinks from authoritative sites, topical depth across a domain, freshness signals, and clear entity association between your brand and the topics you want to own. Everything E-E-A-T rewards. What no longer moves the needle: exact match anchor keyword targeting, thin content built for volume, and generic ultimate guide listicles that summarize other people’s research.

Why LLM SEO Matters More Than Ranking Position Three

Every SEO agency has renamed its offering in the past year. LLM SEO, GEO (Generative Engine Optimization), and AEO (Answer Engine Optimization) all refer to roughly the same discipline: getting your content cited inside AI generated answers across ChatGPT, Perplexity, Google AI Mode, Gemini, and Claude.

Ignore the acronym wars. The underlying question is: when a buyer asks an AI for a recommendation, does your brand show up? Here is what we tell clients when we start this work.

Platforms retrieve differently. Per Virayo’s aggregated retrieval analysis, ChatGPT overlaps heavily with Bing’s index (roughly 87 percent match on top ten results), so Bing Webmaster Tools is not optional anymore. Perplexity leans heavily on Reddit and independent research (46.7 percent of citations from Reddit). Google AI Mode pulls from Google’s own organic index (76.1 percent of citations from top ten pages). Getting cited on all four requires presence across all their retrieval surfaces.

Structure beats volume. In the audits we run, the biggest gap is usually not content quantity. It is content structure. Pages that answer clear questions in clear language get cited. Pages that meander do not. Most SEO content written in the last five years fails this test.

Entity clarity is real. If your brand is not tightly associated with the topics you want to own, LLMs will surface competitors even when your content is better. Building entity clarity is a slow, boring, effective process: consistent naming across the web, structured data markup, expert bylines, and third party mentions that establish who you are.

Almost no client we work with had audited citation rates across major LLMs before we started. Tools like Peec.ai and Profound automate this at scale. Even manual monthly checks give you enough signal to know where to invest. For AI enabled content and LLM SEO strategy, our Agentic Marketing Solution and HubSpot Content Hub practices map the roadmap.

FAST FACT: ChatGPT matches Bing’s top 10 results 87% of the time. Perplexity draws 46.7% of its citations from Reddit. Google AI Overviews pull 76.1% of cited URLs from pages already ranking in the top 10 organic results. Source: Virayo LLM SEO B2B Guide, 2026. 

Not sure if your brand shows up in AI search?

Most SEO programs never audit whether they get cited in ChatGPT, Perplexity, or Google AI Overviews. In two minutes, our RevOps AI CRM Audit surfaces the visibility gaps in your content, data, and technical foundation. No sales call required to see the output.

Start Your Free RevOps AI CRM Audit  → 

What We Actually Do When We Rebuild a Client’s SEO Program

Every rebuild we run follows roughly the same sequence. This is what our Analysis-First methodology looks like applied to organic acquisition.

First we audit the current state honestly. Ranked keywords is not the interesting metric anymore. We look at share of AI Overview citations for target queries, share of voice inside ChatGPT and Perplexity, technical health, structured data coverage, and topical depth versus competitors. Almost every audit surfaces a gap the client did not know existed. Usually the gap is not what they expected.

Then we restructure the top of the funnel. Pages that answer specific questions in the first two paragraphs, with question format H2 headings, get more citations. We usually rewrite the top five or ten highest traffic pages first because the ROI is fastest there.

We fix the technical foundation. Schema markup, especially FAQ and Article and Organization schema, is table stakes now. An llms.txt file is a small signal but a real one. Site speed still matters, mobile first indexing still matters, and internal linking becomes more important because LLMs use it to understand entity relationships. Our Marketing Automation practice handles the CMS side of this work across HubSpot and Salesforce environments.

We build the citation strategy. This is what most SEO agencies still call link building but the discipline is broader now. Digital PR to earn brand mentions in trusted publications, structured comparison content that AI models can cite directly, and third party review presence on G2, TrustRadius, or category specific directories. All of this feeds LLM confidence in your entity.

Finally we set up measurement that reflects the new reality. Not just ranked keywords and organic sessions. AI citation share, brand search volume, conversion rate by traffic segment, and pipeline attribution by AI referral source. If your CFO cannot see how organic contributes to pipeline in your dashboard, the reporting layer needs to be rebuilt.

The Technical Foundations Nobody Wants to Fix

The unglamorous part of SEO in 2026 is exactly what it was in 2023: technical health that everyone knows they should improve and nobody prioritizes. What has changed is that ignoring it is now much more expensive.

Site speed still gates both rankings and conversion. If your page takes longer than three seconds to load on mobile, you are losing traffic no ranking improvement will recover. This has not changed. The tolerance for slow sites has only shrunk.

Structured data has moved from nice to have to necessary. Google’s AI systems and third party LLMs actively parse schema to build their answers. FAQ schema, HowTo schema, Product schema where relevant, Organization schema for brand entity clarity. Missing schema is missing citations.

llms.txt is a newer signal we now add to every client site. It is a small file, similar to robots.txt, that tells AI crawlers what content is available and how to interpret it. Adoption is still low. Adding it is a small effort with a real upside as the standard becomes more widely honored.

Crawler access for AI bots. Many clients accidentally block GPTBot, ClaudeBot, or PerplexityBot in their robots.txt. If you want to be cited by these systems, they need to be allowed to crawl. This is the fastest fix we make on some engagements and it produces measurable citation gains within weeks.

None of this is exciting. All of it changes how much visibility you get. Our Business Process Consultation practice handles the operational side of getting these fixes prioritized when internal teams are stretched.

Measuring SEO When Half Your Traffic Never Clicks

The traditional SEO dashboard is now misleading. Ranked keywords, organic sessions, and CTR do not tell you the story anymore. Here is what we push clients to report on instead.

AI citation share for your target queries across at least ChatGPT, Perplexity, Gemini, and Google AI Mode. Manual monthly audits work if the query set is small. Peec.ai or Profound scale it.

Brand search volume trend. If AI is answering informational queries and your brand appears in the answer, the follow up is often a branded search. Rising branded search on Google, Bing, or LinkedIn is often the leading indicator of SEO working in the new environment.

Conversion rate by traffic segment. AI Overview referred traffic, AI referred traffic (ChatGPT, Perplexity), and traditional organic all now convert at different rates. Reporting on them separately gives you a truer picture of what is working. In most B2B clients we see AI referred traffic converting two to three times higher than average organic.

Zero click impressions. If Google is answering the question in an AI Overview using your content, that is a brand exposure event even without a click. Search Console impressions still matter. Comparing them against click through data now tells you how much of your visibility is happening at the answer layer.

For B2B especially, first touch and last touch attribution both undercount SEO in this environment. We usually push clients toward blended attribution that credits both discovery (which SEO increasingly owns) and conversion. Our Insight and Analytics practice builds the reporting layer for teams making this shift.

What We Have Killed From the Playbook

Here is what we stopped recommending in the past 18 months.

Aggressive keyword targeting on every page. Google understands topics, not keywords. Writing three pages that target three slight variations of the same query cannibalizes traffic and does nothing for rankings.

Thin content built for volume. Any page under 800 words that summarizes existing web knowledge is now a liability, not an asset. LLMs and Google both penalize it. If your content strategy is producing pages of that shape at scale, the strategy is the problem.

Voice search optimization as a separate strategy. Voice search is now just search. Optimizing content to answer natural language questions covers both.

Backlink acquisition as the primary lever. Backlinks still matter, especially from authoritative and topically relevant sites. Guest posts on low authority sites and PBN links do not. If you are still buying links, stop.

Generic ultimate guide listicles. They were the workhorse of 2020 SEO. They do not get cited by LLMs and they do not rank in an AI Overview environment. The signal has moved to specific, authoritative, expert content that stakes a real position.

Optimizing only for Google. ChatGPT alone drives around 87 percent of AI referral traffic in most B2B categories. Ignoring it is leaving a channel on the table. Perplexity and Gemini are smaller today but growing faster than Google’s core search.

Summary

If we had to distill what has changed since our December 2023 SEO post, it comes down to this: SEO is no longer a single surface discipline. Google now shows AI Overviews on close to half of all queries, ChatGPT and Perplexity mediate a growing share of B2B research, and the click through economics of ranking number one have shifted permanently. Rankings still matter, backlinks still matter, technical health still matters. But content that is not structured for AI extraction, not present in the LLM retrieval layer, and not tracked with a modern measurement framework is invisible in ways the old dashboards will not show you.

The teams outperforming their peers have stopped optimizing only for Google’s ranked results and started optimizing for two surfaces at once. That shift requires structural rewrites, entity clarity work, LLM citation monitoring, and a reporting redesign that reflects fewer visitors with higher intent. It is not glamorous. It is what works. If your organic traffic looks worse than your pipeline says it should, that is often the signal that your SEO is doing better than your dashboard shows.

Ready to rebuild your SEO for the AI search era?

Mountainise is a San Francisco based RevOps and marketing consultancy running SEO and LLM SEO programs for enterprise and mid market B2B teams. Our Analysis-First methodology and Lucrative RevOps Framework are how we deliver on organic acquisition, content strategy, and revenue attribution. If your organic traffic is declining while the pipeline holds steady, that is usually a measurement problem, not a strategy problem.

Book a Strategy Session with Mountainise  →

Frequently Asked Questions