When Word to PDF is the right call

Sending a contract to a counterparty for signature. Submitting a CV to a job application portal. Distributing a finalised report where you don't want recipients editing the text. Printing a fixed-layout document where pagination, fonts, and spacing must be identical on every machine. Archiving a final version of a document for compliance. For all of these, PDF is correct — it's the de facto standard for "this is the locked-down final version".

When Word to PDF is the wrong call

Feeding the document to ChatGPT, Claude, or Gemini — PDF is the worst common format for LLM input. Text extraction from PDF is lossy (column ordering breaks, tables get scrambled, headers/footers leak into body text), and you pay roughly 2-3x the tokens versus Markdown for the same content. If the next step in your workflow is "paste this into an AI", convert to Markdown instead. PDF to Markdown exists for the case where you only have a PDF, but starting from .docx, going Word → Markdown is the direct path.

Quality of the PDF output

Layout fidelity is high — fonts, spacing, page breaks, headers/footers, tables, embedded images all render identically to how Word displays the document on screen. Hyperlinks are preserved as clickable PDF annotations. Document metadata (title, author, keywords) is embedded in the PDF. Output is standard PDF/A-compliant for archive use.