> ## Documentation Index
> Fetch the complete documentation index at: https://docs.getnexor.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Agent settings
This page is the reference for configuring an agent: the settings that control who it is, how it talks, which channels it uses, what data it collects, and how it works toward your goal. You reach it from the agent's configuration screen, where each section below appears as its own block. Set the **Goal** first, then align the prompt, pipeline, variables, and knowledge sources with that outcome.
## Goal
The **Goal** setting defines the result the agent pursues in each conversation. It is more than a label: it changes the built-in tools, suggested stages, primary metric, and configuration sections that are available.
The picker offers these goals:
| Goal | `goal_type` | When to use it | Related configuration |
| ------------------- | --------------------- | ---------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------ |
| Book a Meeting | `appointment` | Qualify, check availability, and book a meeting. | Shows **Meeting** and enables calendar tools. |
| Send Payment Link | `payment_link` | Take the conversation through payment and share a checkout link. | Define when to offer it and which tool or integration provides the link. |
| Lead Qualification | `qualification` | Collect facts and decide whether a lead meets your criteria. | Define required variables and pipeline transition rules. |
| Provide Information | `information` | Answer questions about products, services, or processes. | Prioritize a complete, current knowledge base. |
| Customer Support | `support` | Resolve issues and escalate cases that need a person. | Configure limits, escalation, and knowledge sources. |
| Document Collection | `document_collection` | Request and verify the documents needed for a process. | Shows **Documents**, where each required document is defined. |
| Generate a Quote | `quote` | Build and send a quote from the assigned catalog. | Shows **Quote**, where you select a product or service catalog. |
| Custom | `custom` | Pursue an outcome outside the presets, such as debt collection, win-back, post-sale follow-up, or surveys. | Adds an operator-written objective of up to 600 characters. Does not enable booking. |
### Custom goal
Selecting **Custom** reveals **What should this agent achieve?** Describe an observable outcome, not a sequence of instructions. For example: `Recover the overdue instalment with a concrete payment date agreed, without damaging the customer relationship.`
This text is stored as `config.goal_statement` and takes priority over generic preset language in the agent's behavior. Business rules still define the context, limits, and operating instructions.
A `custom` agent does not receive the built-in calendar tools. If the outcome ends in a booked meeting, use `appointment` and express the qualification criteria in the business rules and pipeline.
When the goal changes, Nexor reconciles the goal target in the pipeline and shows the corresponding configuration sections. Review stages, required variables, and tools before processing more leads.
## Your agent
The **Your agent** section defines your agent's identity: the name it uses, who it represents, and how it sounds. These settings shape the agent's tone and how it introduces itself to every lead.
| Field | What it sets |
| ------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Name | The agent's display name, used when it introduces itself (e.g. `Jarvis`). |
| Role / Title | The agent's job title (e.g. `Sales Advisor`). |
| Company | The business the agent represents (e.g. `Stark Industries`). |
| Language | The language the agent speaks, including regional variant (e.g. `Spanish (Chile)`). |
| Time zone | The agent's time zone (e.g. `America/Santiago (UTC-3/-4)`), used for scheduling meetings and respecting contact hours. |
| Voice | The voice used on voice channels and calls. You can filter voices available for the language, preview them, and, when enabled for your account, clone a voice. |
Set the Name, Role, and Company to match how you want leads to perceive your business. The agent uses all three when it introduces itself at the start of a conversation.
## Business rules
The **Business rules** section is your agent's prompt: it describes how your business operates and what the agent should do in each scenario. This is where you teach it to collect relevant information, handle pushback, recognize that the goal has been achieved, and decide when to hand off to a human.
The prompt is split into 7 sections you switch between from a dropdown. Edit one at a time:
| Section | What it covers |
| --------------------- | ---------------------------------------------------------------------------------- |
| Business context | What your business does, whom it serves, and what outcome the agent should achieve |
| Agent limits | What the agent must never do or say |
| Objection handling | How to respond to common pushback and doubts |
| Escalation | When and how to hand off to a human |
| Example conversations | Sample dialogues that model the right tone and flow |
| Additional rules | Any extra instructions that don't fit the other sections |
| Discard rules | Conditions for stopping pursuit or closing the case without achieving the goal |
Each section has its own purpose and writing style. For detailed guidance and examples on filling out each one, see the [Prompt best practices guide](/docs/en/guides/agents/prompt-best-practices).
**Token budget.** A counter shows how much of your prompt budget (around 3,000 tokens) you've used across all 7 sections combined. If you exceed it, tighten your wording rather than removing whole sections.
**History.** Every saved change is versioned. Open **History** to review previous versions and revert to any of them.
**Train your agent.** After editing your rules, click **Train your agent** to apply the changes. The agent uses the updated prompt only after training completes.
Don't try to fill every section perfectly on the first pass. Start with **Business context**, train, and test in the playground, then refine the other sections based on how the agent actually responds.
## Channels
The **Channels** section defines how your agent reaches and replies to leads. Each channel shows a connection status of **Active** or **Connect**. From **Manage**, you can also set channel-specific instructions and style; supported channels allow a model override, and WhatsApp/email manage their templates here.
The **First-contact channel** dropdown sets which channel your agent uses for the first outbound message (for example, WhatsApp).
| Channel | Notes |
| --------- | ----------------------------------------------------------------------------------------------------------------------- |
| WhatsApp | Sends and receives messages from your connected number. See [Connecting WhatsApp](/docs/en/guides/connecting-whatsapp). |
| Instagram | One account per agent. Connect through Meta. |
| Email | Replies from your sender address (for example, `sales@starkindustries.com`). |
| Phone | A phone number used for voice. |
| SMS | A number used for text messages. |
| Web Chat | An inbound widget embedded on your website; it is not available as a first-contact channel. |
Set the **First-contact channel** to the channel your leads are most likely to respond on. WhatsApp is the most common choice.
## Contact hours
Set the windows during which your agent may initiate or resume outbound contact. Incoming replies are handled even outside these windows.
Under **Outside these hours**, **Block calls** prevents calls but still allows WhatsApp, email, or SMS re-contact. **Block all channels** postpones every outbound re-contact until the next window.
Choose which days the agent can reach out, then set the time range for those days.
| Field | Description |
| ---------- | -------------------------------------------------------------------------------------------- |
| Days | When the agent can initiate contact: **Mon-Fri**, **Mon-Sat**, **Every day**, or **Custom**. |
| Start time | The earliest the agent will reach out (e.g. 8:00 AM). |
| End time | The latest the agent will reach out (e.g. 7:30 PM). |
You can add multiple time ranges within a day, for example to pause during a midday break.
With **Custom**, you can define different ranges for each day. Every time is interpreted in the agent's configured time zone.
## Re-engagement attempts
The **Re-engagement attempts** section controls how persistently your agent follows up with a lead who stops replying, until it achieves the goal outcome or exhausts the attempts.
| Setting | Example | Meaning |
| ---------------- | ------- | ---------------------------------------------------------------------- |
| Total attempts | 5 | Maximum number of times the agent re-contacts a lead before giving up. |
| First wait | 3h | Hours to wait after the lead goes quiet before the first re-contact. |
| Between attempts | 24h | Minimum hours between each subsequent re-contact. |
With the values above, the agent re-contacts the lead up to 5 times. The first follow-up goes out 3 hours after the lead stops replying, and each later attempt is spaced at least 24 hours apart. If the lead still hasn't replied after all 5 attempts, they are marked **Lost**.
Under **Advanced**, you can enable rules for eligible middle stages, write per-stage AI instructions, and override the first wait, interval, and maximum attempts. Initial, terminal, paused, and already-achieved goal stages are not offered for re-engagement.
## Variables
The **Variables** section defines the fields your agent collects from each lead during the conversation. As the agent talks, it extracts these values from what the lead says and stores them on the lead record.
Each variable has these properties:
| Property | Description |
| ------------------------- | --------------------------------------------------------------------------------------------------------- |
| **Name** | What you are capturing, for example "Main pain point". |
| **Type** | `Text`, `Number`, `Date`, `Select`, or `Yes/No`. `Select` variables include a closed list of options. |
| **Required** | Whether the agent must capture the value. Optional variables are filled only when the lead provides them. |
| **AI extraction hints** | Signals and formats that help the AI recognize the value correctly. |
| **Prefill from Metadata** | Maps incoming metadata. When the key arrives, Nexor fills the value and the AI does not ask for it again. |
Below are common variables used to qualify a lead:
| Variable | Type |
| ---------------------- | ------ |
| Main pain point | Text |
| Number of users | Number |
| Property type | Text |
| Next committed contact | Date |
| Work zone | Text |
| Interest level | Select |
| Objections raised | Text |
| Current platform | Text |
| Region | Text |
| Time in industry | Text |
Variables power qualification and pipeline transitions: a stage can require certain variables to be filled before a lead advances, so the agent knows what it still needs to ask.
Use **+ Add** to create your own variables. Mark a variable as required when a downstream stage depends on it, and use `Number` for values you want to filter or sort on later.
## Pipeline
The pipeline defines the stages each lead moves through, from first contact to a closed outcome. As your agent talks to a lead, it moves them between stages based on the transition instructions you set.
### Initial stages by goal
Every pipeline begins with system stages for **New**, **Contacted**, and **In conversation**. The winning outcome changes with the goal:
| Goal | Stage that represents the goal |
| --------------------- | ----------------------------------------------------- |
| `appointment` | Meeting scheduled |
| `payment_link` | Payment link sent |
| `qualification` | Qualified; requires the required variables |
| `information` | Information delivered |
| `support` | Resolved |
| `document_collection` | Documents approved, after a paused human-review stage |
| `quote` | Quote sent |
| `custom` | Goal achieved |
You can also add **custom stages** to match your sales process, such as a "Demo taker" stage for leads who have gone through a product demo.
### Terminal stages
Terminal stages mark the end of the pipeline:
| Stage | Type | Meaning |
| ----- | ------------------ | ------------------------------------------------------ |
| Won | Successful outcome | Goal achieved; its concrete label depends on the goal. |
| Lost | Lost | Lead declined or went unresponsive. |
### Configuring a stage
Each stage has its own settings. Select a stage to edit the following:
The stage name shown across the agent and pipeline views.
A color used to identify the stage at a glance.
A short explanation of what the stage represents.
Defines whether the stage is normal, triggered by first contact or first reply, acts as a qualification gate, or represents a booked meeting. Terminal stages are categorized as won, lost, or paused.
Plain-language rules describing when a lead should move into this stage. Reference collected variables with `@`, for example `@budget` or `@company_size`.
Optionally move a lead onward when they go quiet. Set a wait time in hours, then choose the stage the lead moves to once that time passes.
A strict gate that blocks a lead from entering the stage until the required variables are filled. Use this to make sure key information is captured before a lead advances.
Use transition instructions and required variables together: the instructions tell the agent when to advance a lead, while required variables guarantee the data behind that decision is complete.
## Quote
This section only appears when the goal is **Generate a Quote** (`quote`).
Select the catalog from which the agent searches products or services and builds the quote. Catalog content is managed on the global **Catalogs** page; this section only assigns it to the agent. Before activation, test that items include enough names, prices, and attributes to distinguish them.
## Documents
This section only appears when the goal is **Document Collection** (`document_collection`).
Add one requirement per document. Each one defines a name, a unique `snake_case` key, a team-facing description, instructions for how the AI should request it, whether it is required, and accepted formats. You can accept any file or restrict it to JPEG, PNG, PDF, HEIC, WebP, DOC, and DOCX.
Once all required documents arrive, the lead moves to **Documents collected** and the agent pauses for human review. Human approval completes the goal in **Documents approved**.
## Meeting booking
This section only appears when the goal is **Book a Meeting** (`appointment`).
The **Meeting booking** section controls how the agent books meetings once a lead is qualified, what the calendar event looks like, and who attends.
### Event details
Set the title and length of the event the agent creates when it books a meeting.
| Field | Description |
| ----------- | -------------------------------------------------------------------------------------- |
| Event title | The calendar event name, e.g. `Meet Stark Industries`. Supports variables (see below). |
| Duration | Length of the meeting in minutes, e.g. `30`. |
You can use these variables in the event title:
| Variable | Resolves to |
| ----------------- | -------------------------------- |
| `{{leadName}}` | The lead's name |
| `{{hostName}}` | The assigned team member |
| `{{clientName}}` | Your company name |
| `{{meetingType}}` | The type of meeting being booked |
### Booking rules
**Allow overbooking** lets multiple meetings land in the same time slot. Leave it off to limit each slot to a single booking.
### Team member assignment
Choose how the agent picks who hosts each meeting.
* **Selection criteria** — a free-text rule for round-robin assignment. Describe how meetings should be distributed, for example: distribute meetings evenly across the team, assign to whoever currently has the fewest, or always respect each member's availability.
* **Team members** — the list of people who attend meetings, each with their own availability. The agent only assigns slots that fall within a member's availability.
* **Use your own assignment tool** — bypass the built-in logic and let an external tool decide who is assigned.
### Always invite
Add people who should be invited to every meeting, regardless of who is assigned as host. Useful for managers or note-takers who join every call.
### Confirmation email
Toggle on **booking confirmation email** to automatically send the lead a confirmation after a meeting is scheduled.
## Reminders & alerts
Configure the messages your agent sends around each meeting, and the automatic actions it takes when meeting events occur.
### Reminders
Reminders are scheduled messages sent before or after a meeting. Add as many as you need, each with its own timing, channel, and template.
| Field | Description |
| -------- | ----------------------------------------------------------------------------------------------------------------- |
| Timing | When the reminder is sent relative to the meeting, for example `5 min before`, `1 hour before`, or `immediately`. |
| Channel | How the reminder is delivered: WhatsApp or Voice call. |
| Template | The approved message template to use, for example `utility_recordatorio_reunion_v1`. |
Add multiple reminders to build a sequence, such as one an hour before and another five minutes before the meeting.
### Alerts & automatic follow-ups
Alerts and follow-ups are triggered by meeting events. Enable each one independently.
| Event | Action |
| --------------------------------- | ---------------------------------------------------------------- |
| New meeting | Notifies the host that a meeting has been booked. |
| Cancellation (host alert) | Notifies the host by email when a meeting is cancelled. |
| No-show | Automatically re-engages leads who did not attend their meeting. |
| Cancellation (lead re-engagement) | Automatically re-engages leads who cancelled their meeting. |
Reminders use approved message templates. Make sure the template you select is approved on the channel before relying on it.
## Knowledge base
The knowledge base is everything your agent can draw on to answer questions during a conversation. Connect the sources that describe your business, products, and policies so the agent responds with accurate, up-to-date information.
| Source | What it is |
| -------- | ---------------------------------------------------------------------------------------------------- |
| Files | Documents you upload (PDFs, spreadsheets, and other files) for the agent to reference. |
| Websites | Pages from a website. Use a single page or crawl all pages of a site, such as `starkindustries.com`. |
| Catalogs | Structured product or property catalogs the agent can pull listings and details from. |
Use **Manage** to add, edit, or remove sources in each category.
The agent only uses the sources you enable. Anything you leave disabled is ignored, even if it is still listed here.
For setup steps and best practices, see the [knowledge base guide](/docs/en/guides/agents/knowledge-base).
## Agent-level advanced settings
The **Advanced** block inside an agent controls capabilities scoped to that agent. It is different from the global **Advanced** hub in the sidebar, where Webhooks, Background Jobs, Pre-processors, and code functions are managed. See the [Advanced hub guide](/docs/en/guides/advanced/overview) for those global automations.
### Roll up reporting under
Use **Roll up reporting under** to group this agent's numbers with other agents on the dashboard. Pick a group (the master agent), and this agent's metrics are reported together with the rest of that group instead of standing alone.
Group several agents under one master agent when you want a single combined view, for example one figure per team or per region rather than per individual agent.
### MCP tools
MCP tools are HTTP functions the AI can invoke during a conversation, for example to query a CRM, check inventory, calculate a price, or take an action while talking to a lead. Assign them from the shared catalog or create a tool scoped to this agent.
New agents receive a goal-specific tool set. Every goal includes essential functions for saving fields, updating metadata, and moving the lead through the pipeline. `appointment` adds availability, booking, cancellation, rescheduling, and guest tools; `document_collection` adds document request and acknowledgement tools; and catalog-backed goals can search and read products or services. Custom tools extend that set.
| Setting | What it does |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ |
| Pre-agent hooks | Tools that run the moment a lead enters the agent, before any message is sent. Use them to prepare or enrich the conversation up front. |
| Per-stage tool restriction | Limit a tool to specific stages. Select the stages where the tool is allowed; with no stage selected, the tool is available in all stages. |
| Stage actions | Run a tool automatically when a lead enters a stage, without relying on the agent to call it during the conversation. |
A pre-agent hook and a stage restriction are mutually exclusive: the hook runs before a conversational stage can enforce the gate. Stage actions are also automatic and must be given any arguments that the agent would normally write at call time.