The Library AI Actually Cites

Problem

The nonprofit had genuinely useful educational material, but most of it was effectively invisible. More than 200 PDFs were scattered across dozens of website pages, old resource sections, direct file links, and previous email newsletters. Some documents could only be found if someone already had the exact URL, leaving the organization with a large knowledge base but no usable knowledge architecture.

Search engines were not effectively indexing the material, and visitors had no consistent way to browse resources by topic, language, or format. Accessibility added another layer of risk: a formal audit identified dozens of WCAG 2.1 AA violations, creating a serious concern for an organization dependent on government grant funding.

The bilingual audience was also underserved. Spanish content relied on a translation widget that did not create independently indexable pages and could not reliably preserve technical terminology.

Individually, these issues were manageable, but together they limited discoverability, accessibility, credibility, and the organization’s ability to demonstrate digital quality to funders.

Challenge

The nonprofit needed more than a cleaner file library — it needed a structured, searchable resource system supporting:

• Free and paid materials

• Membership-based access

• Filters by topic, language, and format

• Bilingual, indexable content

• WCAG 2.1 AA accessibility

• SEO-friendly pages around key PDFs

• Grant-ready documentation

• Simple publishing workflows for internal staff

The content also had to be easier for search engines and generative AI systems to understand through structured resource pages with clear context and metadata.