The Lead Gen Agent is an omnichannel conversational agent that captures and qualifies leads across web, text, email, and social channels. It automatically detects user intent and sentiment, generates context-aware replies (using configurable prompts or your knowledge base), and can escalate complex conversations to human agents when needed. The agent also supports lead capture, configurable follow-ups, and business-hours rules, and offers a built-in preview/test state so you can validate behavior before going live.
 

Key Capabilities:

• Omni-channel triggers: enable the agent on web, text, email, Facebook, Instagram, and more.
• Intent & sentiment classification: understand user purpose and tone to personalize responses and trigger human handoffs.
• Response Composer: generate replies via custom messages, knowledge-base answers, or prompts (with channel-specific optimization).
• Lead capture & auto-follow: collect lead details and re-engage visitors who drop off via configured follow-up messages.
• Human handoff & notifications: notify team members in real time and provide fallback messaging when humans are unavailable.

Navigation

Go to ‘Settings,’ click ‘AI agents,’ and select ‘Lead Generation agent.’

Goals

The agent automatically engages incoming visitors and prospects, identifies their intent, captures lead information, and intelligently routes conversations to the right destination — whether that’s a human agent or an automated response.

This section outlines the core objectives of the Lead Gen Agent and the measurable outcomes it drives for your business:

Capture more leads – Automatically collect key visitor details such as name, email, phone number, and other custom information fields to convert interactions into actionable leads.

Save time – Reduce manual effort by enabling the agent to instantly respond to frequently asked questions using your connected knowledge base.

Deflect conversations – Let the agent autonomously qualify conversations, resolve simple queries, and escalate complex cases to human agents when required.

Trigger

Use the 'Trigger' tab to define when and where your Lead Generation Agent should run. This ensures the agent responds to messages across selected communication channels and timeframes.

Select communication channels

Click the dropdown next to all channels and select the channels where the agent should respond. Click ‘Apply’ after selecting.

Choose when the agent runs

Set the time window for the agent to respond. Click the time dropdown (e.g., 'Any Time (24/7)' and choose the preferred availability option from the list.

• At any time (24/7) – Agent stays active round-the-clock.
• During business hours – Operates only when your business is open.
• Outside business hours – Engages visitors after hours, ensuring leads are still captured.

Tasks

Switch to the ‘Tasks’ tab and configure the tasks the agent will perform.

Understand and analyze the intent of incoming messages

This task enables the agent to interpret what a visitor is saying by detecting intent (the purpose of their message) and sentiment (tone of the message). This allows the agent to respond contextually and escalate only when necessary.

Click ‘Intent classifier.’ This opens the Intent classifier panel, where you can view and manage the list of default and custom intents.

Default Intents

You’ll find four default intent categories already available - Conversation starter, Appointment, Talk to human, and Unclassified.

These default intents can’t be renamed, but their descriptions can be customized for clarity. Hover over the intent name and click the pencil icon to modify the description.

Adding Custom Intents

You can create new intents specific to your business. 

Click 'Add intent' (top-right corner).

Enter the Intent name and description, and define what this intent represents. Click 'Save' to add the new intent.

The newly created intent can be viewed under the ‘Custom’ section.

The Sentiment classifier helps the agent determine the tone or emotion behind a visitor’s message — such as whether the visitor is happy, neutral, upset, or unclear.

This enables the agent to personalize its tone, decide when to escalate, or deliver empathetic responses automatically.

Click ‘Sentiment classifier.’

This opens the Sentiment Classifier panel, showing four predefined sentiment categories - Positive, Neutral, Negative, and Unclassified. 

To customize a sentiment, hover over the sentiment and click the edit (pencil) icon. Modify the text to include your custom criteria or keywords that define that tone for your business context.

Intent-based response to the message

Generate a reply based on the intent and sentiment, and enhance the reply with specific configurations and CTAs related to each channel using ‘Response composer.’

Select 'Response generator.' This determines how the agent replies based on detected intent and sentiment.

​​Once your intents and sentiments are configured in the previous task, the Response Generator automatically maps them to suitable reply actions. Each predefined intent or sentiment (like Conversation Starter, Appointment, Unclassified, or Negative Sentiment) has its own rule that defines how the agent will respond.

In this section, you can only view or edit these pre-configured response rules — you won’t see an option to add a new one unless you click 'Add response' manually. Hover over the response generator and click the edit icon.

When you click 'Edit' for Sentiment is Negative, you’ll notice the 'Custom message' option is selected. This means a specific message was already configured in a previous setup. This predefined message is automatically triggered whenever the agent detects a negative sentiment in a customer’s message.

If needed, you can change how the agent replies by switching the action type in the dropdown.

Available options include:

Custom message – Manually written response shown to the visitor.

AI answer using knowledge base – Let the AI pull an answer from your business’s knowledge base.

AI prompt – Uses AI to dynamically generate a response based on detected intent and sentiment.

Add a Custom Response Rule

Click ‘Add response’ in the top-right corner of the Response Generator screen.

Under ‘Conditions,’ set up when the rule should trigger.

Click the first dropdown and you can select either ‘Intent’ or ‘Sentiment.’ For example, select ‘Intent.’  From the second dropdown, choose the intent (for example, Insurance & Cost). 

All other intent options are disabled since they’ve already been used in previous response setups, leaving ‘Insurance & Cost’ as the only available choice.

Under 'Actions,' choose how the agent should respond.

Then choose one of the following options:

Custom message – Create your own response text directly in the editor.

AI prompt - When you select AI prompt as the action type, the agent uses generative AI to craft a dynamic and context-aware response instead of pulling from a fixed message or knowledge base entry.

• You can guide how the AI should respond by entering a prompt in the message field — for example, specifying tone, content type, or instructions like “Respond politely with brief pricing details and invite the visitor to share contact information.”
• The 'Personalize' option allows you to insert placeholders (like customer name, business name, and more) to make the message more conversational and tailored.

AI answer using knowledge base – Allow the AI to fetch an accurate, data-backed answer from your organization’s knowledge base.

Clicking the arrow corresponding to updating the knowledge base content will redirect you to the ‘Knowledge’ section, where you can add or bulk import FAQs related to your business.

Once the intent and action are defined, click ‘Save.’

Select ‘Response optimizer’ to tailor your AI agent’s responses for each communication channel.

In the ‘Web’ tab, you can customize how the agent appears and interacts with visitors on your website.

Reply Buttons

Set up quick-response buttons that visitors can use during conversations:

Resolution Button – Displays when a visitor confirms their issue is resolved. Example: “That helped 👍”

Escalation Button – Allows the visitor to request human assistance. Example: “Talk to Human 🧑‍💼”

Click ‘Show preview’

View how these buttons will appear in real time.

Customize AI Agent Name

Choose how your AI assistant is introduced to visitors.

• Enter a name (e.g., Jarvis) that appears in chats and helps humanize the conversation.
• Use 'Show Preview' to see how the name appears in live chat.

Fallback Message

This message appears when the AI agent cannot find an answer.

During business hours: Display a message that directs users to the available support team.

After business hours: Let users know that the team is offline and assure them of follow-up.

Both responses can include personalized placeholders such as business contact details.

To turn off the fallback message, toggle the switch next to Fallback message to the off position.

Once disabled, the AI agent will no longer display any fallback messages when it cannot answer a question.

Click ‘Save’ in the top right corner after configuration.

Switch to the ‘Text’ tab. It lets you configure how your AI agent manages SMS or text-based interactions.

Add Unsubscribe Text

Enable this option to include an unsubscribe message in your outgoing texts.

• This allows customers to opt out of receiving further SMS communication.

Fallback Message

This section defines what message the AI agent sends when it cannot provide an answer via text.

You also preview how ‘During and After business hours’ messages appear in the text form.

Click ‘Save’ to apply the configuration.

The ‘Email’ tab helps configure how the AI agent handles fallback messages for email conversations.

Fallback Message

This section defines the response your AI agent sends when it cannot answer a visitor’s email query for ‘During’ and ‘After’ business hours.

The ‘Social’ tab lets you configure how your AI agent interacts with users on connected social media channels such as Facebook and Instagram.

Channels

This shows the connected social platforms where the AI agent can respond — for example, Facebook and Instagram.

Social Responses

Indicates that the AI agent’s responses on social platforms will be sent without pre-defined templates.

Fallback Message

Specifies what message the AI agent sends when it cannot provide an answer on social channels.

• During business hours: Directs customers to contact the team for assistance.
• After business hours: Acknowledges the query and promises a follow-up once the team is back online.

‘Contact us’ form - Based on your selection, the AI agent will use the preferred outreach channel to contact customers. It will follow the existing configuration for that channel.

Human handoff for complex needs

The Escalation Notifier is used when a customer’s intent is to speak with a human agent. It helps automate the handoff process by sending appropriate messages and notifications to both customers and business users.

Click ‘Escalation Notifier.’

Under the ‘Web’ tab, you can view:

During Business Hours

Send Default Message

The AI agent sends a confirmation that a human colleague is being notified and may take a few minutes to join.

Send Notification to Business Users

Sends an internal alert notifying team members that a new message has been received from a visitor.

Send Fallback Message (after 3 minutes) 

If human agents haven’t responded in time, the AI sends a follow-up message asking if it can share details via text or email instead.

Show Closing Message (after 3 minutes of inactivity)

Ends the chat politely by providing an alternate contact option.

Outside Business Hours

Show Offline Message

Automatically informs the visitor that the business is currently closed and their message has been received.

Show Closing Message

Wraps up the interaction with a friendly note and a reminder to reach out via the preferred channel.

Similarly, you can view and configure notifier messages for Text, Email, and Social channels. Each tab follows the same structure as the Web tab — displaying default, fallback, and offline messages that can be previewed.

Capture Lead  

The 'Capture Lead' feature allows the AI agent to collect visitor details such as name, phone number, email, and other required fields during a conversation to generate leads.

Web 

Capture Leads

Define when and how the agent should collect customer information.

Before chat starts – Collect lead info before initiating the chat.

During chat – Collect details while chatting (recommended).

Do not capture leads – Skip lead collection entirely.

Lead Capture Fields

Mandatory fields are chosen and included in the form.

Preview and Save

You can preview how the lead capture form will appear to visitors and click Save to finalize the configuration.

Text

The Text tab allows you to create leads using only the mobile number.

When enabled, leads are automatically created without requiring the visitor’s name or email ID. Additional details can still be appended later during the conversation.

Email

The Email tab enables lead creation using only the email ID.

When this option is turned on, leads will be generated without requiring a mobile number. Additional contact details can later be appended as the conversation progresses.

Social 

The Social tab lets the AI agent create leads from connected social platforms like Facebook and Instagram.

• Leads are created using only the visitor’s name.
• The system does not require an email or mobile number initially — those can be appended later during the conversation.

Contact Us Form 

The Contact us form option allows leads to be created when either an email ID or a mobile number is provided.

This ensures that even minimal contact information can generate a valid lead for follow-up.

Auto-follow up

The Auto-follow-up feature re-engages visitors who leave the chat before completing their conversation.
Using the Visitor follow-up tool, you can configure how and when to send reminder messages.

In the 'Web' tab, you can:

• Enable or disable the follow-up option.
• Set automated follow-ups after the visitor has been inactive for 3 minutes.
• Customize different messages for during business hours and outside business hours (both include placeholders like [Business Name] and [Business Phone]).

This ensures that potential leads are not lost and conversations can continue later.

Text, Email, and Social

These sections show that Text-based follow-ups are not yet supported.

A note indicates that support for follow-up messages will be available soon.

Preview and Testing the Agent

You can preview how the Lead Generation Agent responds to visitor queries directly from the Tasks tab.

Select 'Start a test chat' to begin a simulated conversation with the agent.

In the example shown:

The visitor asks for details about insurance and cost.

The agent replies with a detailed, informative response explaining coverage, copays, and cost comparisons between urgent care and emergency visits.

It also provides source links for further reading. Finally, the agent requests the visitor’s name to proceed with lead capture and offers quick reply buttons like “That helped 👍” and “Talk to Human 🧑‍💼”.

After reviewing the results, if you’re satisfied, pause the agent first.

After pausing the agent, the ‘Publish’ button becomes visible. Click it to publish the agent.

Outcomes

The Outcomes section tracks how effectively the Lead generation agent performs against its defined goals.

Leads captured – Displays the total number of leads collected across all channels, including anonymous ones.

Deflection rate – Shows how many conversations were successfully handled by the AI without triggering a Talk to Human request.

Time saved – Reflects the cumulative time saved through AI responses across all incoming messages, also showing the equivalent cost savings.

You can also filter results for different time ranges (e.g., Past 30 days) to analyze performance trends.

Configuring Time Saved

You can fine-tune how Time saved is calculated by selecting the customize icon next to the metric.

A pop-up window lets you define:

• Average time saved by the agent for each message handled
• Hourly employee wage in your preferred currency

Once updated, these values automatically adjust the cost saved figure displayed under the Time Saved section — helping quantify the financial impact of automation.