> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zelinqa.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Glossary

> Definitions of key terms used in the NBQ Engine documentation.

| Term                     | Meaning                                                                                                                                                                                                   |
| ------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **NBQ**                  | *Next Best Question* — the core logic that selects the single most useful question to ask next in a conversation.                                                                                         |
| **NBQ Engine**           | The service that runs the NBQ logic, exposed as a REST API.                                                                                                                                               |
| **NBQ instance / agent** | One configured agent within your tenant. It has its own question bank, behaviour settings, tone, and API keys. A tenant can have several instances.                                                       |
| **Tenant**               | Your organisation. All your data and agents live under one tenant. Requests are scoped to it automatically via your API key — you never pass a tenant identifier yourself.                                |
| **Session**              | A single conversation, identified by your `session_id`. Usage is counted per session. The first call with a new `session_id` registers the conversation; subsequent calls with the same ID belong to it.  |
| **Category**             | A group of questions that collect the same piece of information (e.g. budget, timeline, use case). Categories can carry a priority order.                                                                 |
| **Conversion**           | A recorded outcome of a conversation (`purchase`, `lead`, `abandon`, or `other`), reported via the conversion endpoint and used for NBQ Studio analytics.                                                 |
| **exhausted**            | The API field that signals there are no more questions worth asking. When `true`, `next_question` is `null`.                                                                                              |
| **external\_id**         | The stable identifier of a question, set in the NBQ Studio. Feed it back into `answered_question_ids` once asked so NBQ never repeats it.                                                                 |
| **Pattern A**            | Integration pattern where `next_question.text` is displayed directly to the user. Simplest approach.                                                                                                      |
| **Pattern B**            | Integration pattern where `next_question.text` is passed to your own LLM or agent, which rephrases it naturally inside its own reply.                                                                     |
| **context**              | A free-text API field (≤ 8000 chars) summarising the conversation so far. Lighter-weight alternative to sending full `conversation_history`. Also where user facts go — there is no `user_profile` field. |