Docs/AI agents
Prompt vs rule vs scenario vs knowledge
Every AI setting in Sendify answers a different question. Once you know which question you are answering, it is obvious where your instruction belongs.
On this page
The building blocks
| Building block | The question it answers | Where to configure it |
|---|---|---|
| Prompt | Who is the AI and how should it behave? | Prompt editor |
| Rule | What must it always or never do? | Rules |
| Scenario | When a specific situation happens, what should happen next? | Scenarios |
| Knowledge | What factual information can the AI use? | Knowledge base |
| Canonical fact | Which fact wins when sources disagree? | Canonical facts |
| Qualification question | What information must be collected? | Qualification |
| Phase | What stage of the conversation are we in? | Phases |
Which one do I use?
If you catch yourself writing "it should never say" or "it must always ask", that is a rule, not prompt text. If you are writing "when the lead says they are busy", that is a scenario. If you are writing a price, an address, or an opening time, that is knowledge, and if it must beat anything else the AI has picked up, make it a canonical fact.
Keep the prompt about character
Go deeper
Every block above is documented in full inside the AI builder guide. This page exists so you know which section to open.
Related articles
Still need a hand?
If your volume is going to run past the daily or hourly limits, or something is not behaving the way these docs describe, talk to our team.