The new playbook for organic growth in the AI search era

The old organic playbook still works. Just less than it used to.

If you have been watching Search Console for the last twelve months, you have probably seen the same shape we have. Impressions holding steady. Average position stable. Click-through rate falling off a cliff. Your content is still being seen. People are just getting the answer somewhere above your link.

That somewhere is the AI Overview, the ChatGPT answer, the Perplexity citation, the Gemini summary. Search has become a layer of synthesis sitting on top of the index, and the funnel has moved upstream of the click. The page you optimised for is still there. It is just no longer the destination.

This post is about what changed, why the old assumptions broke, and what to actually do about it as a founder who cannot afford a six-person SEO team.

What changed in 2026

Three shifts happened in parallel. All of them are measurable.

Clicks are leaving the page

Seer Interactive analysed 3,119 informational queries across 42 organisations between June 2024 and September 2025. On queries where Google showed an AI Overview, organic click-through rate dropped from 1.76% to 0.61%. That is a 61% decline. Paid CTR on the same queries dropped 68%.

By February 2026, ALM Corp tracked the recovery. CTR on AI Overview queries crept back from 1.3% in December to 2.4% in February. Better than the bottom. Nowhere near what it was. The floor is not falling further, but the ceiling is permanently lower.

Zero-click is the new default

60% of all Google searches now end without a click. On mobile, that figure is 77%. In Google’s newer AI Mode, 93% of searches never produce a click at all. The user types a question, gets an answer, and the conversation is done. No website was harmed in the making of this answer.

The brands cited in AI answers no longer match the brands ranking in Google

This one is the quietest and the most disruptive. In mid-2025, around 75% of pages cited in AI Overviews also ranked in the top 10 organic results. By February 2026, that overlap had collapsed to between 17% and 38%, depending on the study.

BrightEdge tracked a 400% increase in AI citations pulled from pages ranking in positions 21 through 30. 89% of AI citations now come from sites ranked beyond the top 100 organic results entirely. Whatever AI engines are using to decide what to cite, it is not your Google ranking.

If you have internalised this stack, you already know the old goals (rank in the top 3, capture the click, drive traffic) no longer add up to the same outcome. Ranking is now necessary but not sufficient. There is a second filter after ranking, and that is where most founders are getting eaten.

Why traditional SEO assumptions broke

This is not "SEO is dead." Search is still where most B2B buyers start. People still type questions into Google. The rules around what happens after they type have changed, and the rules favour a different mix of signals than the ones we spent a decade optimising for.

Ranking is now necessary, not sufficient

The old game was simple. You ranked, you got the click. The conversion happened on your page. Search engines were a routing layer between intent and content.

The new game adds a second step. After you rank, an AI engine reads the SERP, decides what to synthesise, and chooses which sources to cite by name. Sometimes it cites you. Sometimes it cites three sources and you are number 47. Sometimes it answers without citing anyone. Whether you appear in the citation set is a separate problem from whether you rank.

You can rank second for a query and not be cited in the AI Overview sitting above you. Your competitor can rank fourteenth and be cited. That is the new reality. The buyer reads the AI summary, sees your competitor’s name, and never scrolls down to your link.

AI engines weight different signals than Google

Google ranks pages. AI engines weight evidence. Different mechanism, different inputs.

When Perplexity decides what to cite, it crawls 8 to 12 pages per query in real time and selects 3 to 4 to absorb into the answer. Selection depends on how cleanly the page makes specific, citable claims. Pages full of hedged language and vague statements get dropped even if they rank well. Pages with concrete numbers, bounded claims, and clean structure get cited even from low-authority domains.

ChatGPT works differently. It leans heavily on its training data, supplemented by web search when the user asks for current information or the model judges it necessary. The training data was scraped at a specific cutoff. If your brand was not in the corpus at scrape time, you are effectively invisible to ChatGPT for that release, no matter how good your SEO is right now.

Gemini and Google AI Overviews share infrastructure with Google search but apply different ranking on top. The most striking finding: only 11% of sites are cited by both ChatGPT and Perplexity. Platform fragmentation is real. Optimising for one engine does not carry over to the others.

This is why "AI SEO" advice often contradicts itself. It depends which engine you are trying to be cited in.

The compound-mention model

Here is the underlying pattern that makes sense of the data. AI engines look for agreement across sources. When ChatGPT or Perplexity decides whether to recommend a product, they check whether the recommendation appears across multiple independent sources. Your own website does not count as one of them. It is the baseline. Everything beyond that is the signal.

A brand mentioned 200 times across Reddit threads, three review sites, two podcast transcripts, and a Hacker News discussion is treated very differently from a brand mentioned only on its own domain. Even if both have identical SEO. The first has compound authority across the corpus. The second has self-published claims. AI engines weight the first heavily and dismiss the second as marketing.

This explains why scrappy brands sometimes punch wildly above their weight in AI citations and why some well-funded brands with great SEO get completely ignored. The compound-mention surface is what gets cited. SEO is the cost of entry. Citation density across the wider web is the differentiator.

What works now

We will not list ten things. We will list three.

Map your channels before you optimise anything

Most founders are still treating SEO as "publish on our blog, build backlinks, wait." That works for some categories. It does not work for most.

Where AI engines source from depends entirely on what your buyers research. A category like "best CRM for SaaS startups" gets pulled heavily from Reddit, G2, and indie blogs. A category like "enterprise data warehouse benchmarks" gets pulled from analyst reports, vendor docs, and Gartner. Same engine, completely different source mix.

Before you write a single blog post, sit down for an hour and map this out. Pick five queries your buyer would type. Run them through ChatGPT, Perplexity, and Google AI Overviews. Look at the citation patterns. See where the answers actually come from. You will often find the bulk comes from places you were not planning to publish on.

For B2B SaaS specifically in 2026, the source mix that keeps showing up is Reddit threads, G2 and Capterra reviews, Hacker News discussions, and a handful of high-authority blogs. Your own blog matters, but as a citable reference for claims made elsewhere, not as the primary discovery surface.

Build authority across surfaces, not just on your own site

This is the part most founders find hardest because it requires showing up in places that do not feel like marketing.

The mechanism is simple. Pick four or five places your buyers gather. Show up there consistently for months. Not with promotional posts. With useful comments, helpful answers, occasional product mentions when relevant. Build a recognisable presence. The goal is not immediate traffic. It is getting your name into the corpus that AI engines read.

Reddit is the highest-leverage single channel in 2026 for most B2B SaaS. One Superprompt analysis found Reddit appears in 68% of AI answers across the queries they tracked. It is the most-cited single source on the open web. Reddit hit 1.4 billion monthly visits by April 2025 and AI citations from Reddit grew 450% between March and June 2025. If your buyers hang out on Reddit, this is where the compounding happens.

But Reddit is not always right. SparkToro is worth using before you commit to any channel. Sometimes the leverage is on Hacker News. Sometimes it is a niche Discord. Sometimes LinkedIn, if your buyer is enterprise. The point is not Reddit specifically. The point is: pick the channels deliberately, then show up enough that AI engines treat you as part of the conversation in your category.

Measure the right signals

Most marketing dashboards still measure things that no longer correlate with growth. Rankings, traffic, CTR. These were leading indicators for revenue when search was a routing layer. They are not anymore.

The metrics that survive 2026 are different:

• Citation rate. The percentage of buyer-relevant AI prompts where your brand appears as a cited source. For B2B SaaS, below 10% means you are invisible. The healthy band is 20% to 30%. Anything above 40% is category leadership. Seed-stage startups realistically start at 2% to 8% and build from there.
• Share of voice across engines. Not just whether you are cited, but how often versus your competitors when buyers ask category questions.
• Sentiment of mentions. When ChatGPT describes your product, is the description accurate, outdated, flattering, or wrong? Inaccurate descriptions are common and surprisingly recoverable once you know they exist.
• Prompt coverage. Of the 20 to 50 queries your buyers actually run, how many produce your brand in the answer set? This is where you discover the gaps that your blog traffic does not show you.

None of this shows up in Google Analytics. You need to actually run the queries against the engines, or use a tool that does it for you.

What this means for solo founders

The headline reality is uncomfortable. The signals AI engines weight are exactly the signals solo founders are worst at producing. We do not have PR teams to land mentions in TechCrunch. We do not have community managers running our Reddit presence. We do not have research teams producing the kind of structured, citable content that gets absorbed cleanly into AI answers.

But the headline reality also flips. The signals AI engines weight are signals that no amount of money can quickly buy. Compound mentions across genuine community discussions take years to build. Enterprises with full marketing budgets often score worse on AI citation than scrappy founders who have spent two years being genuinely helpful in three subreddits. The 2026 game rewards consistent presence in conversations more than it rewards budget.

What this means in practice is that the work is doable but it has to be done. Not delegated to an agency that publishes 12 blog posts a month. Not solved by a one-time content sprint. The compounding happens in the small daily acts of showing up where your buyers are, contributing something useful, and being mentioned by name when someone else describes the problem you solve.

Most founders will not do this. Most growth tools will not help with it because most growth tools are still measuring last decade’s metrics. The founders who will win the AI search era are the ones who treat AI visibility as a daily practice instead of a quarterly project.

That is the playbook. It is not complicated. It is just different from what worked five years ago, and most of the advice on the internet has not caught up.

Where to go from here

If this resonated, two things are worth doing next.

First, audit your own AI visibility. Most founders have never run the basic diagnostic on their own product. The next post in this series walks through how to do it manually in fifteen minutes, or in sixty seconds with the free AfterLaunch snapshot.

Second, understand the specific best practices and pitfalls that determine whether AI engines cite you or ignore you. The third post in this series covers what we have seen work and what we have seen fail across hundreds of snapshot scans.

Or join the waitlist for early access to AfterLaunch, the always-on growth engine that runs this discoverability work for you across every surface.