Services
Let us show you where we and how can contribute to your product literature publishing workflow.
| Workflow Step / How we can help | Use of Large Language Models / Artificial Intelligence 🤖 |
|---|---|
| 1. Legacy content conversion This is usually the most complex and challenging step in every project: We will examine your existing product literature (manuals, quick guides, brochures) and give you a quote for converting this legacy base to modern, web-based documents. Depending on the number and complexity of documents and the source formats, this may take anywhere from a few days to months. In general, PDF conversion is more expensive and time-consuming than working from open (editable) source formats such as Microsoft Word and Adobe InDesign. |
Use of AI Artificial intelligence/large language models can be used to structure content scraped from PDFs or other legacy formats, but humans will always be required to evaluate and post-edit such “AI-cleaned” documents. |
| 2. Technical writing While we do not employ technical writers, we can help you make your content web-ready and translation-ready. This includes formatting cleanup and making content translation-friendly. Making a document (set) translation-friendly usually involves sentence-level edits where we will remove ambiguity and passive voice, and fragments are converted into full sentences. |
Use of AI Large language models can be used for making content more translation-friendly, but humans should evaluate and post-edit LLM output. |
| 3. Typesetting and structuring content Depending on the quality of the source documents, no or little formatting may be required at this step. Plain, unformatted content, however, needs to be formatted/structured. We use a set of dedicated scripts and macros in this step. |
Use of AI Generic large language models are usually not a great help when it comes to formatting content. |
| 4. Adding metadata Metadata such as dates, product categories and document types make documents more useful and allow us to fine-tune the publishing process as well – for example, metadata can be used to present or hide certain template sections. We use a set of dedicated scripts and macros in this step. We can also process prepared metadata from brands. |
Use of AI Large language models can be used to create metadata “seeds” ; for example keywords derived from a text. |
| 5. Translation In general, there are three options for translation: (1) We can have your content translated by professional human translators; (2) We can publish machine-translated content as such with a custom degree of editing (from “as is” to meticulous proofreading and post-editing), or (3) We can publish your content in English only, relying on browser-/user-side machine translation, which we support with a customizable language drop-down. |
Use of AI In all three models, AI is used: (1) Professional translators will use machine translation, translation memories (databases with previously translated content) an termbases to deliver solid translations. (2) Post-editing machine translation content is a common workflow for combining the best of both worlds (quick machine translation) and human expertise for ensuring solid results. (3) Modern browsers have access to machine translation engines, both as a built-in feature and using extensions. |
| 6. Publishing Finished documents are converted / “rendered” as HTML documents and optionally as PDFs. We can publish documents for you, or you can use our web interface for triggering the process. |
Use of AI No AI is used in this step. |
Investing in a modern documentation publishing format and workflow will lay a foundation for many years to come. We are ready to bring your brand’s manuals, quick guides and FAQs into the 21st century.
↻ 2024-08-05