Knowledge Base
The support knowledge the AI Service Copilot reads from, so agents can answer customers consistently and faster.
Knowledge Base
The knowledge base is the library of support content the AI Service Copilot reads from when it drafts responses, suggests resolutions, or triages cases. A well-maintained knowledge base is what makes the Copilot useful — it's the difference between vague AI fluff and accurate, on-brand customer answers.
What's in the knowledge base
The Support Knowledge index covers four content types:
| Content type | Examples |
|---|---|
| FAQs | "How do I reset my password?", "What are your support hours?" |
| Troubleshooting guides | Step-by-step problem-solving for common issues |
| SLA policies | The promises you make to customers — response and resolution times |
| Resolution templates | Reusable answer drafts for recurring case types |
These are stored as Markdown files under /knowledge/support/ and re-indexed automatically when changed.
How the Service Copilot uses it
When an agent asks the Service Copilot to "suggest a resolution" or "help me reply to this customer", the Copilot:
- Takes the case subject and description as the question.
- Searches the knowledge base for the 5 most relevant snippets above a quality threshold.
- Drafts a customer-facing response grounded in those snippets.
- Cites the source for each claim, so the agent can verify before sending.
This means the AI never makes things up — it can only repeat what's in your knowledge base.
How the Sales Copilot uses other knowledge
HotCRM ships with four AI knowledge bases, each tuned for different teams:
| Knowledge Base | Used by | Content |
|---|---|---|
| Sales Knowledge | Sales Copilot | Sales playbook, qualification methodology, objection handling |
| Product Information | Sales Copilot | Product catalog, specifications, pricing, positioning |
| Support Knowledge | Service Copilot | FAQs, troubleshooting, SLA policies |
| Competitive Intelligence | Sales Copilot | Competitor profiles, battlecards, win/loss analysis |
See AI Copilot › Knowledge Bases for the technical details.
Curating the knowledge base
The knowledge base is only as good as what you put in it. A typical content-curation workflow:
1. Capture what's working
When an agent resolves a tricky case with a great answer, copy that answer into the knowledge base. The next agent (or the AI) will reuse it.
2. Triage incoming questions
Every week, look at the most common case subjects. The top 10 questions deserve dedicated FAQ entries — if 30 customers asked the same thing, the 31st should be answered automatically.
3. Retire stale content
If a product changes, the corresponding KB articles need updating. Set a quarterly review on each article's owner.
4. Watch the AI's citations
When the Service Copilot drafts a response, it cites which KB articles it used. If the same article is cited every day, it's pulling its weight. If an article is never cited, consider rewriting or removing it.
Tips for service managers
- ✅ Make KB updates part of the weekly team meeting — "what 3 new articles do we need this week?"
- ✅ Reward agents who contribute KB content; it scales their expertise.
- ✅ Audit the AI's recent answers — pick 10 at random per week and check the cited articles were used correctly.
Tips for admins
- Place new Markdown files under
/knowledge/support/and they'll be indexed on the next embedding refresh (configurable interval). - The AI uses retrieval augmentation — it never trains on your KB, just searches it at query time. New content takes effect immediately after the index updates.
- For details on chunking, embeddings, retrieval thresholds and caching, see AI Copilot › Knowledge Bases.
- You can add new knowledge bases (e.g., a separate compliance KB) and wire them to specific skills. See Customization › Adding AI Skills.