Expert Knowledge Capture

Captures employee know-how from internal documents and expert review so it can be reused in knowledge work.

Business impact

• answer accuracy — Nissan’s article says the system’s answer accuracy improved during use.
• documentation modeling time — Honda’s IBM case study says the pilot was expected to slash documentation modeling time by 67%, though the source notes this was a pilot rather than a measured production result.

Data requirements

• presentation materials (text) — Honda’s pilot extracted know-how from PowerPoint materials.
• diagrams (image) — Honda’s source materials included diagrams.
• graphs (image) — Honda’s source materials included graphs.
• internal documents (text) — Nissan’s system answers questions using internal documents.
• accumulated know-how — Nissan’s application also uses know-how contributed over time.

AI methods and techniques

• generative AI — Used at Honda to extract know-how from PowerPoint materials, including diagrams and graphs, for knowledge modeling.
• information extraction — Used to pull reusable knowledge out of internal documents and presentation materials.
• question answering — Used in Nissan’s internal application to answer employee questions from stored knowledge.
• human-in-the-loop review — Experienced employees review responses and add knowledge so the system improves over time.

AI models and model families

IBM watsonx.ai

Industries

Real-world evidence

2 documented case studies on record.

Companies using this: Honda, Nissan.