Gumshoe-generated content is built around two goals: getting crawled and read by AI models in the first place, and then persuading the model to actually recommend your brand once it has read you. We weight the second goal, influence, more heavily than most tools do, since being findable but unconvincing still leaves you out of the answer.
How is the content structured so models can chunk it?
Content is broken into stand-alone sections that each make sense on their own, since models often process a page in pieces rather than all at once. Each section opens with a short context sentence and closes with a takeaway, and we avoid cross-references like "see above" that only make sense if a reader has the whole page in front of them.
On top of that, content follows a clear heading hierarchy: one H1 that summarizes the page, H2s for each major section, and H3s for sub-points. Lists and steps are broken into bullets or numbered items, one idea per line, kept to a handful of items so they stay scannable for models and people alike.
Why does the content include so many citations and statistics?
This is deliberate, not a quality issue. Gumshoe content is sculpted for an AI audience, and evidence like citations, statistics, and expert quotes is part of what makes content easier for a model to trust and cite. If a piece reads as more formal or more heavily referenced than typical marketing copy, that is by design.
How is the content actually written?
It is grounded in your own Visibility Audit (formerly known as a report): your specific prompts and personas, not a generic prompt. Generation runs through a multi-step pipeline: models research the topic, a set of models drafts the content, and then a separate set of models reviews and critiques that draft before it's finalized. That pipeline is continuously refined against a large, growing base of real citation data, so it reflects what has actually gotten brands cited, not just a guess at good writing.
Once your content is ready, you can export it as HTML, a Word doc, Markdown, or plain text, whichever fits how you publish.
What about the technical side, like schema and structure?
Great content still needs to be crawlable. Your Technical Audit checks whether your pages use structured data (schema.org markup) and clean, semantic HTML, and whether AI models can actually crawl and read them at all. It gives you a prioritized fix list so you know exactly which technical issue to tackle first, rather than a long list of everything that could be better.
Pro tip: if a piece of content isn't showing up in AI answers, check your Technical Audit before assuming the content itself is the problem. Even well-optimized content can go unread if a model can't crawl the page it's published on.