From the firm’s own archive to citeable content — across web, social, and media

A 30-year Caribbean firm already had decades of expertise sitting in the wrong shapes: legacy site pages that crawlers couldn’t parse cleanly, a long archive of educational lessons and commentary, practice-area copy that was thin or uneven, and video explainers that lived on YouTube without a matching written asset. The gap wasn’t “buy a chatbot.” It was: turn what the firm already owned into a reusable content knowledge base, then ship it in the formats search and answer engines — and clients — actually use.

I built the pipeline around that archive.

First, inventory and structure. We pulled the firm’s existing materials into a working knowledge base: topic maps of what had already been taught, statute and judgment anchors that belonged with each subject, practice-area gaps vs. what clients actually search, and which videos or older lessons could become evergreen articles. AI helped cluster themes, extract candidate questions, draft outlines, and spot where two old pieces were really one topic — so humans weren’t starting from a blank prompt every time. Nothing privileged went into a consumer model; this layer was the firm’s public and reusable expertise, organized so every new piece could cite sources and stay jurisdiction-specific.

Second, SEO and AEO as one writing standard. Every long-form piece was designed answer-first: the question a client actually asks, then the firm’s position, then the statute, case, or primary source. That is how search snippets and answer engines lift from directories — and how a practice stays citeable instead of generic. Technical SEO sat underneath: clear entity and service structure, clean URLs, citation-ready metadata and schema on pages meant to be quoted, redirects so years of equity survived the rebuild, and templates so a practice page, a bio, and an article were distinguishable to people and machines. AI accelerated research scaffolding and first drafts against the knowledge base; lawyers and editors kept information architecture, legal positioning, factual review, and the publish button.

Third, publish once, derivate many. The same approved article became more than a website URL. From the knowledge base and the long-form piece we produced: site articles and landing pages; shorter posts for the firm’s channels; social-ready captions and link posts with proper titles and preview images; image assets for cards and previews; and audio/video paths — either turning a spoken explainer into a written insight, or spinning a written piece into shareable clips and companion posts. Social wasn’t an afterthought: link previews, featured images, and refresh workflows so a share pulled the right title and art instead of a blank cache. The point of the system is one approved unit of knowledge, many surfaces — web library, posts, image, audio, video, social — without rewriting the law from scratch each time.

What this is not. It is not a claim that the firm “shows up in ChatGPT.” It is not unsupervised AI publishing. It is not the private brief bank or the babysat private copilot — those are separate later Work cards. This card is the outward content machine: firm data in, structured knowledge base, SEO/AEO articles out, then the same truth pushed across websites, posts, and media.

Status: knowledge-base-backed SEO/AEO publishing system delivered for the public sites and multi-format distribution. Humans still approve anything that goes live.