For a decade, the playbook for software companies was stable. Rank for the right keywords, earn the click, convert the visitor. In 2026 the first two steps are coming apart. Google answers a growing share of questions on the results page itself, and buyers increasingly ask ChatGPT, Perplexity or Gemini for a vendor shortlist before they open a search engine at all.
NeuroSEO, a European boutique SaaS SEO agency working with English-speaking software and AI companies, has spent the past year rebuilding its approach around that shift. The agency’s position is unusual in a market that tends toward extremes: neither “SEO is dead” nor “nothing has changed”, but a specific argument about which parts of search still compound and which have quietly stopped working.
The click is disappearing. The citation is what is left
The numbers behind the shift are no longer anecdotal. Seer Interactive’s analysis of 3,119 queries and 25.1 million impressions in September 2025 found that organic click-through rates fell by roughly 61% when an AI Overview was present on the results page. Advanced Web Ranking reported that AI Overviews appeared on close to half of tracked queries by the fourth quarter of 2025.
On the buyer side, 6sense’s research on B2B purchasing found that 94% of buyers use large language models at some point in the buying process. For software specifically, that process is long, comparative and increasingly conducted without the vendor’s knowledge. By the time a demo request arrives, the shortlist has usually been formed, and a model may well have helped form it.
“The question we hear from SaaS founders is whether SEO still makes sense,” says Vasiliy Rusakov, Head of SEO at NeuroSEO. “It does, but the unit of success has moved. Ranking first for a query nobody clicks is not a result. Being the vendor a model names when a buyer asks for options is.”
Why software companies feel it first
SaaS products break several assumptions that generic search advice is built on. There is no catalogue to expand and no local intent to capture. The queries with real commercial value are often the smallest ones: a comparison between two tools, an alternative to a market leader, a specific integration. A few hundred searches a month from the right people outweigh fifty thousand from the wrong ones.
Those same bottom-of-funnel queries are exactly where AI answers have landed hardest. A buyer asking “best invoicing software for freelancers” no longer needs to visit three pages. The assistant summarises them and names a handful of products. If a vendor is absent from that answer, no ranking position fixes it, because the buyer never saw a results page.
This is why the agency argues that SaaS companies need to treat classic search and generative search as one programme with two outputs rather than two separate budgets. The technical and content groundwork is shared. What differs is what gets measured.
What a specialist agency does differently
NeuroSEO describes its SaaS work as a sequence rather than a list of tactics. According to the agency, the order matters more than the individual pieces, and getting it wrong is the most common reason retainers underperform on subscription products.
- Fix access before anything else. If a crawler cannot render the page, nothing downstream works. On SaaS sites this is usually a JavaScript problem rather than a content problem.
- Start where intent is. Comparison, alternatives and integration pages produce measurable pipeline within months. Top-of-funnel guides come later, once there is authority to carry them.
- Write comparisons honestly. Pages that admit where a competitor fits better convert more reliably, and they are the ones AI systems quote. A page that pretends the competition has no strengths is treated as marketing by readers and models alike.
- Keep the company’s facts consistent everywhere. Models assemble a picture of a vendor from its site, review platforms, LinkedIn, Crunchbase and the trade press. When those disagree, the model picks one or blends them.
- Report against signups and trials, not sessions. Traffic is an input. The agency says it would rather lose a reporting argument than a client’s quarter.
The agency’s pricing is also notable in context. Specialist SaaS SEO retainers in the United States are commonly quoted at $10,000 to $20,000 per month. NeuroSEO’s full SEO programme starts at 2,500 euros, which it attributes to a senior-only team of around thirty specialists working from Europe rather than a layered account structure.
Timelines follow the same logic. The agency tells prospects to expect bottom-of-funnel pages to produce measurable enquiries within the first few months, while topical authority and citation share build over six to twelve months. The first ninety days go to an audit: technical health, current visibility in both search and AI answers, and the vocabulary buyers actually use. A roadmap follows, sequenced by expected return rather than by what is easiest to publish.
From ranking to being named
The practice of optimising for AI assistants has acquired a name, generative engine optimization, and a certain amount of hype to go with it. NeuroSEO’s view is deliberately unglamorous. Most of what gets a company cited by a model is the same work that gets it ranked. The site has to be crawlable and parseable. Content has to be written in self-contained passages that survive extraction. Structured data should describe the page in machine-readable terms. And the company needs mentions in publications the models already trust.
Where the two diverge is in how success is tracked. The agency runs a fixed set of buying prompts across ChatGPT, Perplexity and Gemini on a recurring schedule and records whether a client is cited, mentioned or absent, and which competitors appear instead. Because assistants answer probabilistically, it reports the trend across runs rather than individual screenshots.
The technical layer most sites miss
One finding the agency returns to repeatedly concerns rendering. Google executes JavaScript before indexing a page. The crawlers behind most AI assistants do not. A crawl-log study published by Vercel with MERJ in December 2024 found that none of the major AI crawlers executed JavaScript: GPTBot and ClaudeBot fetched JavaScript files in 11.5% and 23.8% of their requests respectively, and ran none of them.
For a typical SaaS marketing site built as a single-page application, the consequence is blunt. The product description, the pricing table and the documentation may be invisible to the systems a buyer is asking. The fix is not exotic. Server-side rendering or static generation puts the content back into the initial HTML, where every crawler can read it. But it has to be done, and most sites have not done it.
The agency also points to robots.txt as a frequent self-inflicted wound. Blocking GPTBot keeps content out of OpenAI’s training data, but OAI-SearchBot is a separate crawler that feeds ChatGPT’s search results. Sites that block everything by template default then wonder why assistants never mention them.
Regulation as a demand driver
A second theme in the agency’s recent work is regulation. The European Accessibility Act became enforceable on 28 June 2025 and applies to companies selling to consumers in the EU regardless of where they are registered. NeuroSEO has added accessibility audits and code-level remediation to its services. Its argument is that the semantic markup, labelling and stable layouts assistive technology depends on also serve crawlers. The overlap is real, the agency says, and rarely sold as one piece of work.
The agency has also published a set of industry pages for technical B2B verticals where buyers research heavily before making contact:
- SaaS, AI services and developer-facing software
- Maritime and shipping technology, including emissions-compliance platforms
- Logistics, customs and e-invoicing software under the EU’s ViDA reforms
- RegTech and compliance software for DORA, NIS2 and the AI Act
- Aviation and aerospace, laboratory informatics, and renewable energy software
The common thread, the agency says, is a buyer who reads thoroughly, compares in public and asks machines before asking vendors. That buyer rewards depth over volume, and punishes a single wrong fact in the vocabulary of their industry.
What should not change
For all the emphasis on AI, the agency is careful to say what it is not recommending. It is not advising clients to abandon Google, which still delivers the large majority of organic traffic for software companies. It is not recommending that content be rewritten for machines at the expense of readers; content that reads well to a technical buyer, it argues, is also content a model can quote. And it does not promise rankings or citations, on the grounds that nobody controls either.
“Organic search compounds. Paid stops the moment you stop paying,” Rusakov says. “That has not changed. What has changed is where the compounding happens. Some of it is now inside an assistant’s answer, and most software companies have not started working on that yet.”
About NeuroSEO
NeuroSEO is a boutique search agency serving English-speaking SaaS, AI and software companies across Europe. The team of more than thirty senior specialists handles strategy, content, technical SEO, frontend development and DevOps in-house, with sixteen years of work in search and over two hundred projects delivered. Services include SEO, generative engine optimization, web accessibility and EAA compliance, web development and, through a long-standing partner, app and game development.
