Website Rebuilt for Search and AI Citation: B2B SaaS

A B2B SaaS company was effectively invisible to the way buyers now research. Its website was slow to surface in Google and barely registered with AI assistants like ChatGPT, Claude, Gemini, and Perplexity.

New pages took weeks or months to get indexed and cited, so qualified buyers researching through both search and AI never found the company. I rebuilt the site and its content with a structure optimized for traditional search and AI citation at once, using Claude Code and a focused tooling stack to build and write in parallel. New content earned Google placement and LLM citations within days of launch, and the whole rebuild took roughly 70% less time than a conventional developer-and-writer effort.

The challenge

The site was losing buyers before a conversation ever started, on two fronts at once.

• New pages took weeks or months to get indexed and cited, so fresh content never arrived in time to catch an active buyer.
• The site was structured for neither modern search nor LLM citation, so AI assistants rarely surfaced or quoted it.
• A conventional rebuild, a developer and a writer working in sequence, would have been slow and expensive, and would not have fixed the citation problem on its own.

Buyers now research across both Google and AI assistants. Being absent from either means losing the deal before the company is even in the running, and the old site was effectively absent from both.

The approach

I treated this as a single rebuild aimed at two audiences, the search index and the language models, and built it fast enough that content could keep flowing afterward.

I: Implementation Planning

I defined the dual target up front: rank in search and be citable by LLMs. That set the content structure, the schema, and the indexing approach that serve both, plus a plan to produce the build and the content in parallel rather than in the usual slow sequence. I also decided which concrete, sourced claims the content needed, since those are what language models actually quote.

M: Migration & Execution

I rebuilt the site and its content against that structure, using Claude Code alongside a focused tooling stack so engineering and writing happened together instead of one after the other. The pages were optimized for clean, fast indexing and for the specific, verifiable claims that earn citations.

The results

Indexed and cited in days, not months. New content and structure earned Google placement and LLM citations within days of going live, closing the gap that used to cost the company active buyers.

More qualified traffic and more opportunities. Buyers researching through both search and AI assistants started finding the company, and that traffic converted into more sales conversations.

Roughly 70% less build time. The parallel build-and-write approach delivered the entire rebuild in about 70% less time than a conventional developer-and-writer project.

Why this matters

Buyers research in two places now: search engines and AI assistants. A site built for only one of them misses half the market, and most sites are still built for neither.

The fix is not more pages. It is a structure and a set of concrete, sourced claims that both Google and the language models can read and trust, shipped fast enough that the content never goes stale. Get that right and a site stops being a brochure nobody finds and starts being the thing buyers cite.