Knowledge Base

The Knowledge Base is an internal workspace for reference material. Keva's AI uses active entries when it drafts answers, and staff can search and filter the entries here. Customers do not open the Knowledge Base directly.

What it supports

• Add an entry by typing or pasting content.
• Load one .txt or .md file into the editor.
• Upload one .pdf file for text extraction.
• Import resolved or closed tickets as AI suggestions.
• Add a category label to each entry.
• Edit or archive entries.
• See whether an entry is searchable and whether indexing is pending or has failed.

How it works

1. Add, paste, upload, or import reference material.
2. For a manual or text/Markdown entry, enter its title, content, and category, then save it.
3. Keva queues the entry for indexing so the AI can use it.
4. Search the entries by text or category, and review the indexing status shown on each entry.

PDF uploads create an entry after extraction. Ticket imports create suggested entries for staff review.

AI suggestions

Entries created from resolved or closed tickets appear in AI Suggestions with their title, content, and category. Staff can accept, edit, or dismiss a suggestion. Accepted suggestions become active entries; dismissed suggestions are archived.

Indexing

An entry can be searchable, indexing pending, indexing failed, or not searchable yet. An entry without a semantic index can still be found by keyword matches in its title or content. When the indexing controls are available, staff can retry indexing for eligible entries.

Getting started

1. Open Knowledge Base for the selected workspace.
2. Choose Add to Knowledge Base.
3. Paste or type the content, load a text file, upload a PDF, or select resolved tickets to import.
4. Complete the title and category when the editor asks for them, then save or review the resulting suggestion.
5. Search the list to confirm the entry and check its indexing status.