Search used to behave like a maze. A buyer typed a question at the entrance. They walked a corridor of ten blue links. They then took whichever turn looked most promising. What we called SEO was really one discipline. You learned the maze well enough to stand at the turn your buyer would take.
That maze still exists, but it is about SEO and AEO (AI Search) now. It is no longer only about SEO. For a large share of your buyers and potential customers, it is no longer the maze they walk.
Two Things Happened At The Same Time
The first is that someone now stands at the entrance and answers the question before anyone goes in.
In the first four months of 2026, 68% of US Google searches ended without a click to the open web. That is up from 60% in 2024. The figures come from SparkToro’s analysis of Similarweb clickstream data. Searches that produced at least one click fell from 30.5% to 22% across the same period. Where an AI Overview appears, click-through rates drop by close to 60%.
The buyer asked their question and got an answer. They never entered the maze at all.
The second change is that there is now more than one maze. Google, ChatGPT, Perplexity, Gemini and Claude assemble their answers from different sources, using different logic. They disagree with each other far more than people assume. Loganix synthesised six studies between October 2025 and March 2026. Only 11% of the domains ChatGPT cited also appeared in Perplexity’s citations for the same queries. The same question, put to a different assistant, leads to a different exit. Your business is either named at that exit or it is not.
What is at stake there is not really traffic. It is whether you make the shortlist.
Why This Lands Harder In The B2B World
The same analysis found that 73% of B2B buyers now use AI tools in their purchase research. It also found that buyers substantially complete 61% of purchase decisions before they contact a vendor. Treat any single vendor’s synthesis with appropriate caution. The direction is consistent across every dataset we have seen. It also matches what our clients tell us about their own pipelines.
Put those two numbers together and the implication is uncomfortable. Someone is drawing up the shortlist your sales team competes on, in a conversation you are not part of. The sources are ones you did not choose, about a market you thought you understood. By the time an enquiry reaches you, the field has already narrowed.
For a CEO, that is a commercial risk before it is a marketing one. For an investor, it is a diligence question. The channel itself is quietly changing shape. A business whose inbound pipeline depends on it carries a valuation exposure. That exposure will not reach the management accounts until the year it does.
And Why Hiring One Person No Longer Solves It
When there was one maze with a stable layout, one good person could learn it. Hiring an in-house SEO manager was a sound decision, and for years it was the right one. Plenty of companies went further, and worse, handing marketing, SEO and anything else going to one Marketing Manager.
The work now looks different. Being the answer, rather than a link in the corridor, requires several things to be true at once:
- Technical foundations that let every engine crawl, render and parse the site correctly
- Structured data and entity clarity, so a machine knows what your company is, what it sells and who it serves
- Content that earns a quote rather than only a ranking, which is a different craft with different rules
- Source material worth citing in the first place: original data, clear methodology, named experts
- Third-party presence, because assistants weight what others say about you, not only what you say about yourself
- Measurement across several engines at once, since a position in one tells you very little about the others
- Classical SEO underneath all of it, because Google organic is still the largest single channel for almost every B2B business
That is six or seven disciplines. It is a technical skill set, a content skill set, a data skill set and a PR skill set. One person cannot hold all of them at a senior level. The few who come closest are not taking a single in-house role.
The maze has multiplied and started redrawing itself. You cannot send one person in and expect them to map all of it.
The Cost Side. Facts.
The in-house route is also not the cheap option it appears to be.
Indeed puts the average UK SEO manager salary at £38,358. That figure covers 448 job postings, last updated on 4 September 2026. That is the headline number. It is not the cost.
Add employer National Insurance at 15% on earnings above the £5,000 secondary threshold. On that salary it comes to a shade over £5,000 a year. Add the minimum 3% auto-enrolment pension contribution on qualifying earnings, close to £960. You are at roughly £44,300 before the person has opened a laptop.
Then the tooling. Semrush lists a One Pro+ seat at $299 a month. That lands somewhere around £2,700 to £2,900 a year, depending on the exchange rate. AI visibility tracking is typically a separate line again. Call the total cost of employment somewhere around £47,000 in year one. Add a recruiter’s fee at 15% of salary and you are closer to £53,000.
Being Accurate About The Numbers
Two caveats, because we want to make this argument accurately. The £10,500 Employment Allowance can offset employer NI. It applies to your total NI bill rather than to each hire. An established payroll has usually consumed it already. And a more senior SEO lead in London costs considerably more than the national average above.
So the real comparison is not a £38,000 salary against an agency retainer. It is roughly £50,000 a year, plus your own management time. That buys one person covering one part of a problem that now has six parts.
Full access to a specialist team generally comes in under the fully loaded cost of that single hire. That team covers SEO, AI search, content, technical web and design. The work is not cheaper. You are simply buying the hours you need from each discipline instead of a full-time salary in one.
What The Work Looks Like In Practice
We will not publish client performance figures in a blog post without the client’s name attached. We also do not use numbers we cannot evidence. What we can describe is the shape of the work and where it consistently moves.
Across our client base, which includes software businesses, private equity and M&A firms and portfolio companies, the pattern repeats. We start with an audit of the site and the competitive set. The technical foundations are almost always weaker than the client believes. The competitive picture is almost always different from the one they carry in their heads. We then rebuild the content programme around the questions buyers actually ask. The structure lets a search engine and a language model both lift a clean answer from the page.
We treat schema markup, entity definition and internal linking as part of the content work. They are not an afterthought for a developer. Then we measure visibility across the assistants as well as in Google. A client winning in one and invisible in the others does not have an SEO problem. They have a distribution problem with a very specific fix.
How The Results Arrive
What moves first is usually visibility on the questions that sit closest to a buying decision. What moves next is the quality of the inbound conversation. A buyer who arrived through a clear answer to a real question turns up better informed and further along. Traffic volume is the metric clients ask about first and the one we care about least.
Would you like the specific numbers, under the client’s name, with the reporting behind them? We will walk you through them in a meeting. We would rather show you that properly. Writing a percentage into a blog post gives you no way of checking it.
Where Do You Start?
You do not need a strategy day to find out whether this is a problem for you. Open ChatGPT, Perplexity and Google. Ask each of them the question your best-fit buyer would ask before they know your company exists. Something like “best [your category] software for [your customer type] in the UK”. Read what comes back.
If you are named, look at which source it cited. That page is doing the work, and it may not be yours. If you are not named, look at who is. That list is your live competitive set, and it is frequently not the one on your board pack.
Either way, you will learn more in ten minutes than in a quarter of reporting on keyword positions.
Talk To Us
Has your growth plateaued? Is inbound quieter than it was eighteen months ago? Are you weighing up an in-house hire against outside support? We should talk before you commit to a salary.
Would it be worth a short Teams or Zoom meeting with myself and my co-founder Zafar Shah? We will look at where you currently appear across search and the AI assistants. We will tell you honestly whether we think there is a problem worth solving. Then we will set out what it would take to fix it. If the answer is that you are in good shape, we will tell you that too.
Remora works with founders, CEOs and investors as part of the value-creation team, not as a supplier of marketing packages.
Start the conversation here.
SOURCES
SparkToro, analysis of Similarweb clickstream data, January to April 2026, reported by Search Engine Land https://searchengineland.com/google-zero-click-searches-2026-study-479717
Loganix 2026 B2B AI Buying Behavior Analysis, published 3 April 2026 https://www.prnewswire.co.uk/news-releases/73-of-b2b-buyers-use-ai-tools-in-purchase-research-multi-source-analysis-finds-302733321.html
Indeed UK, SEO Manager salary, 448 salaries, updated 4 September 2026 https://uk.indeed.com/career/seo-manager/salaries
Employer National Insurance rates and thresholds 2026/27 https://employerscalculator.co.uk/guides/employer-ni-rates-2026-27
Auto-enrolment rates and qualifying earnings band 2026/27, Moorepay https://www.moorepay.co.uk/payroll-hr-rates/automatic-enrolment/
Semrush pricing 2026 https://www.demandsage.com/semrush-pricing/
