Search work that starts with whether your site can be crawled, rendered and understood — by Google and increasingly by the AI systems people now ask instead. Rankings follow the technical foundation, not the other way round.
There is no content strategy that rescues a site search engines cannot render, a URL structure that duplicates itself four ways, or a page that takes six seconds to show anything on a phone. We fix the foundation first: crawl budget, indexation, canonicals, redirects, structured data, Core Web Vitals.
A growing share of research now happens inside ChatGPT, Gemini, Claude and Perplexity, which read pages differently from a traditional crawler. They favour clear semantic structure, direct answers near the question, explicit entity information in JSON-LD, and content that states facts plainly rather than building to them.
That is not a separate discipline from good technical SEO — it is the same discipline with the emphasis moved. We structure content so both a crawler and a language model can extract an accurate answer.
Crawl, index, render, duplication, redirects, log review.
Field data diagnosis and the fixes, in priority order.
A connected JSON-LD graph, validated, not scattered snippets.
Pillars, clusters and internal linking with intent.
hreflang, market targeting and duplicate management.
Search Console and Bing, with the AI-citation picture alongside.
Technical fixes can move things in weeks — indexation and Core Web Vitals problems often show improvement inside a month. Competitive ranking gains from content and authority realistically take six months or more. Anyone promising page one in thirty days is selling something else.
Largely by being genuinely citable: clear semantic HTML, direct answers to real questions, accurate structured data describing who you are, and content that exists as text rather than inside images or client-rendered JavaScript. It also helps to be referenced on sites those systems already trust.
No, and neither can anyone else — the ranking systems are not ours to control. We commit to the work, the measurement and honest reporting on what moved.
Usually yes. Most failures come from a handful of causes — unoptimised images, render-blocking resources, third-party scripts and layout shift from unsized elements. We diagnose from field data rather than lab scores, because that is what Google actually uses.