The challenge
The product sits in a category — personalized readings — that is crowded with low-trust, low-quality sites. Search engines and AI systems are rightly skeptical of the whole neighborhood. Standing out required proving, structurally, that this site is different: real sources, real pages, real operator.
The original build was optimized for the product experience, not for retrieval. Key content lived behind interactive flows that crawlers and AI systems could not see, which meant the site's actual substance — readings grounded in citable Torah sources — was invisible to the systems deciding what to recommend.
Trust infrastructure was missing. No privacy policy or terms pages, no branded 404, thin entity signals. These are the boring pages, but both classic search and AI engines use them as legitimacy checks, and their absence quietly caps how far a small site can go.
As a one-person product, distribution could not depend on manual posting. Any content channel had to run without daily attention or it would stop existing.
The strategy
Separate the product from the discoverability layer. The interactive reading experience stays as-is; around it, build a crawlable architecture that exposes the site's real substance — what the readings are, what sources ground them, who runs it — as plain, retrievable pages.
Differentiate on sources. The readings are grounded in verifiable Torah sources, and that grounding is the site's honest advantage over the low-trust competition. The strategy was to make that grounding machine-visible: structured data, entity work, and content that cites what it draws on.
Adopt AI-specific conventions early and cheaply: llms.txt to describe the site to AI crawlers, structured data for the page types that exist, and complete trust pages (privacy, terms, a branded 404) so nothing about the site pattern-matches to the spam in its category.
Automate distribution. Build an Instagram content autopilot that publishes on a schedule without manual intervention, so the site has a steady top-of-funnel presence that doesn't depend on anyone remembering to post.
The execution
Rebuilt the crawlable architecture so the site's substance is reachable by crawlers and AI systems, not locked inside interactive flows. Added llms.txt describing the site for AI crawlers.
Implemented structured data across the site's page types and did the entity work connecting the product, its sources, and its operator into a coherent, machine-readable picture.
Shipped the trust layer: privacy policy, terms of service, and a branded 404 page — the legitimacy signals that separate a real product from the category's noise.
Built and launched the Instagram content autopilot, which now publishes scheduled content without manual intervention, giving the product a continuous social presence feeding the site.