Publish trusted knowledge
Create articles with ownership, visibility, revision, and lifecycle instead of an ungoverned document pile.
Coming soonApproved support knowledge
Knowledge Base for AI Customer Support
Publish trusted help content, project approved Product and Blog revisions into searchable sources, track freshness, and prepare scoped evidence for future AI customer service answers.
- Approved sources remain traceable
- Freshness and revision state are explicit
- Form and CRM PII stays outside retrieval
Knowledge Base
Source and retrieval workspace
Source statusLive concept
Keep support knowledge current and retrievable
Product, Blog, and native documents keep their own domain rules; the Knowledge Base consumes versioned snapshots.
Connect approved sources
Project active Product and published Blog content while preserving original ownership.
Prepare accurate AI retrieval
Index current evidence with source identity, revision, scope, and citation information.
Knowledge Base features
Good AI support depends on a small, current, governed source of truth—not unrestricted access to every workspace record.
Coming soon
Native knowledge documents
Publish support articles with revisions, visibility, scope, and ownership.
Coming soon
Product and Blog adapters
Consume approved snapshots without merging business CRUD into the Knowledge Base.
Coming soon
Source freshness and index health
Show source revision, indexed revision, stale state, failures, and last success.
Coming soon
Scoped access and visibility
Separate public, site-scoped, and internal sources.
Coming soon
Search and retrieval preview
Test a query and inspect which chunks and sources support the result.
Coming soon
Citations and traceability
Carry source identifiers and URLs into support answers and review tools.
How Knowledge Base works
Publish an approved revision, create a stable source snapshot, and replace the retrieval index safely.
- 01
Publish approved content
Create a native document or publish an eligible Product or Blog revision.
- 02
Build and index a snapshot
Normalize, chunk, embed, and atomically replace the current source index.
- 03
Retrieve fresh evidence
Customer service queries only current, permitted sources and preserves citations.
Who Knowledge Base is for
Support customer answers, internal review, help content, and product discovery without indexing private customer data.
Customer help content
Organize setup guides, billing explanations, policies, troubleshooting, and FAQs.
Product answers
Make active product descriptions, specifications, and approved FAQs retrievable.
Support team reference
Give authorized team members one current source of answers and evidence.
AI support grounding
Provide public, fresh, site-scoped evidence for accurate chatbot answers.
Not a warehouse for every workspace record
Only approved content sources enter retrieval. Customer records, submissions, secrets, drafts, and stale content stay outside public AI search.
- Form submissions and attachments are excluded
- CRM notes and sensitive fields are excluded
- Only published or active revisions are indexed
- Archived sources must remove old chunks
Knowledge Base questions
What can enter the knowledge base?
Native knowledge documents and approved projections from Product and Blog are planned first.
Will AI see unpublished content?
Public retrieval must exclude unpublished, archived, stale, expired, or non-public sources.
Does it store CRM customer data?
No. CRM data and form submission PII are not knowledge sources.
When will Knowledge Base be available?
It is planned as a Pro capability and remains in development.
Explore connected website tools
Continue with the most relevant parts of the Cogo Site workflow.