Service · Digital Experience

SEO and AI search for complex offerings.

The relevant search terms in technical B2B are often searched only a few hundred times a month, but by exactly the people who control a budget. At the same time, part of the research is moving into AI assistants.

Start a project See the Growth Architecture

In short

SEO & AI Search consists of three parts: a technically sound website, a dedicated page per search intent and expert content that answers a question in full. For AI assistants, clear definitions, structured data and an unambiguous company entity are added. First movement after about three months, robust results after six to twelve months.

Definition

Generative engine optimisation (GEO) refers to optimising content and site structure so that AI assistants such as ChatGPT, Perplexity or Google's AI Overviews cite a company as a source or name it as a supplier. GEO builds on good SEO and places higher demands on structure, clarity and verifiability.

01 · Starting point

Why standard SEO recipes fall short in technical B2B.

Common SEO advice comes from markets with high search volume and short buying cycles. Capital goods follow different rules.

A term with 90 searches a month can be worth more than one with 9,000 if the 90 searches come from technical buyers with six-figure budgets. What matters is who is searching and at which stage.

In technical markets, the text that actually answers the question wins: with figures, limits, selection criteria and honest caveats. Such content can only be written with expertise from inside the company.

What is new is research via AI assistants. They prefer sources with clear structure, unambiguous definitions and traceable origin.

02 · Approach

How organic visibility is built.

01

Technical audit

1 to 2 weeks
Full crawl of the website: load times, status codes, redirect chains, duplicate content, missing markup, indexing issues.
- Prioritised list of issues
- Load time analysis per page type
- Indexing report
02

Intent analysis

1 to 2 weeks
Collecting the relevant queries from search data, sales conversations and the search terms report of existing ad accounts. The latter is the most valuable and most frequently overlooked source.

Each query is assigned to a stage: problem awareness, solution comparison or supplier selection.
- Topic map with stage assignment
- Mapping of search intent to page
- Gaps compared with competitors
03

Structure and content

ongoing
Building missing pages and revising existing ones. Articles are based on interviews with your engineers.
- New service and industry pages
- Articles with sources
- Definitions and comparison tables
04

AI search and markup

1 to 2 weeks
Structured data for organisation, services, FAQs and articles. Clear entity data, short answers under headings and an llms.txt with a machine-readable profile.
- JSON-LD for all page types
- Entity data and llms.txt
- Checks of mentions in AI assistants
05

Measurement

ongoing
Monthly review: visibility per topic, position of core terms, organic enquiries and their qualification in the CRM.
- Monthly report
- Ranking trend per topic
- Mapping of organic enquiries to orders

03 · AI search

Visibility in ChatGPT, Perplexity and Google AI.

A growing share of B2B research starts in an AI assistant. The requirements overlap strongly with classic SEO, but are stricter in some places.

What makes content citable for AI assistants
Requirement	Implementation	Why
Readable without JavaScript	Server-rendered HTML	Many crawlers do not execute JavaScript
Clear entity	Complete organisation data as JSON-LD	The assistant needs to know who the supplier is
Clear definitions	Definition blocks with `<dfn>`	Short, extractable answers are cited more often
Short answer before long text	Summary directly below the heading	Easier to take over into a generated answer
Verifiability	Date, publisher, sources	Recency and origin are selection criteria
Structured statements	Tables and ordered lists	More reliable to process than running text
Access allowed	robots.txt admits AI crawlers	A blocked crawler cannot recommend you

This website is built on these rules itself, including an llms.txt with a machine-readable profile.

Frequently asked

Frequently asked questions about SEO and AI search.

How long does it take for SEO to work?

Technical improvements take effect within weeks. In technical markets, new content usually needs three to six months to reach stable positions, and robust enquiry numbers appear after six to twelve months. If you need enquiries sooner, run paid search in parallel.

Is SEO worthwhile with low search volume?

Often especially then. If 120 people a month search for a specific type of plant and your average order value is six figures, a position among the top three results is worth a great deal.

Can you guarantee mentions in ChatGPT or Perplexity?

No, and serious providers do not promise that. What you can do is create the conditions: clear entity data, clear statements, structured content and mentions on trustworthy industry sites. We check mentions regularly with a fixed set of test questions.

Do you buy backlinks?

No. Bought links have little effect in technical niches and carry risk. Links come from expert content, industry portals, associations and partners. It takes longer and lasts.

How do you measure success?

We measure visibility per topic, organic enquiries and, once the CRM is connected, how many of them lead to proposals. Rankings are an intermediate step.

Relevant next steps.

Working together

Let's talk about where you stand.

In a 30-minute first conversation we identify where your growth is stuck and whether working together makes sense. You then get an honest view of a realistic scope.

Start a project See resources

We usually reply within 24 hours