Set up your AI agent

Set Up Your AI Agent

The AI agent can reply to customers across connected channels, answer questions using your business knowledge, qualify leads, and help book appointments. Configure the agent's persona, goals, channels, availability, and handoff rules so it behaves consistently with your customer workflow.

What is an AI Agent?

An AI agent is a Vaanexa assistant that communicates with customers on connected channels according to the configuration you provide.

The agent can use your business knowledge to answer customer questions, qualify leads, and support actions such as booking appointments. You can control where and when the agent is active and define situations where a human team member should take over.

One AI agent is created during onboarding, and you can configure it from Agents.

Why use an AI Agent?

An AI agent helps you:

• Respond to customers across connected channels.
• Answer common questions using your business knowledge.
• Qualify leads through conversations.
• Support appointment booking.
• Maintain a consistent communication style.
• Provide responses outside normal business hours when configured.
• Hand conversations to humans when required.
• Reduce repetitive work for your team.

Real-Life Use Case

The situation

Imagine UrbanNest Realty receives customer enquiries throughout the day through WhatsApp, Instagram, and other connected channels. Customers ask about properties, requirements, appointments, and quotations.

The problem

The team cannot respond personally to every enquiry immediately, especially when messages arrive outside business hours.

What happens without the feature?

The team may:

• Delay responses to new enquiries.
• Spend time answering repetitive questions.
• Manually qualify every prospect.
• Miss opportunities when customers need assistance outside working hours.
• Handle different conversations inconsistently.

How Vaanexa helps

UrbanNest configures an AI agent with a specific persona and goal. The agent can answer questions using the business's knowledge, qualify prospects, and help customers move toward an appointment.

Handoff rules ensure that conversations requiring human attention can be transferred to the team.

The new workflow

Customer message → AI response → Knowledge-based answers → Lead qualification → Appointment or next step → Human handoff when required

Case Study

Business

UrbanNest Realty

Challenge

UrbanNest wants to respond quickly to customer enquiries while allowing its sales team to focus on conversations that require direct human involvement.

Objective

Configure an AI agent that can answer common questions, qualify prospects, and support appointment booking while maintaining a clear path to human assistance.

Vaanexa Solution

UrbanNest configures the agent's persona, goal, active channels, operating hours, and handoff rules.

The team adds business knowledge so the agent can answer customer questions accurately. The agent is tested using the built-in test chat before being used with customers.

Outcome

UrbanNest can use AI to handle suitable customer conversations while maintaining human oversight for situations that require a team member.

How the AI Agent works

The AI agent uses several configuration areas to determine how it communicates with customers:

1. Persona defines the agent's name, tone, and language behavior.
2. Goal defines what the agent is expected to accomplish.
3. Channels determine where the agent is active.
4. Hours determine when the agent can operate.
5. Handoff rules define situations where a human should take over.
6. Knowledge provides the information the agent can use to answer customer questions.
7. Approvals can provide additional human review for AI actions.

The agent uses these settings together when handling customer conversations.

Before you begin

Before configuring your AI agent:

• Make sure the required communication channels are connected.
• Decide what the agent should help customers accomplish.
• Prepare accurate business knowledge for the agent.
• Decide when conversations should be handed to a human.
• Determine when the agent should be active.
• Review which AI actions should require approval during initial testing.

How to set up your AI Agent

1. Open Agents.
2. Select your agent. One agent is created during onboarding.
3. Configure the Persona.
4. Set the agent's name, tone, and language.
5. Define the Goal.
6. Choose the Channels where the agent should be active.
7. Configure Hours.
8. Choose whether the agent should always be available or operate only outside business hours.
9. Configure Handoff rules.
10. Define topics or situations where a human must take over.
11. Add relevant business knowledge.
12. Save the configuration.
13. Use the test chat to verify how the agent responds.

Configure the Persona

Set the agent's:

• Name
• Tone
• Language

The agent can reply in the customer's language when it can.

Choose a tone that matches the way your business communicates with customers.

Define the Goal

Specify what you want the agent to accomplish.

Examples include:

• Book a visit.
• Collect customer requirements.
• Share a quotation.
• Qualify a lead.

A clear goal helps the agent focus conversations on useful outcomes rather than simply producing endless text, humanity's favorite form of productivity theater.

Select Channels

Choose the connected channels where the AI agent should be active.

Only enable channels where you want the agent to handle customer conversations.

Configure Hours

Choose when the AI agent should operate.

You can configure it to:

• Stay always on.
• Operate only outside business hours.

Configure Handoff Rules

Define the topics or situations where a human team member should take over.

Examples include:

• The customer requests a person.
• The conversation requires specialized assistance.
• The agent cannot confidently answer the customer's question.
• The conversation matches a configured handoff condition.

How to test your AI Agent

Use the test chat available on the agent screen before relying on the agent with real customers.

1. Open your AI agent.
2. Open the test chat.
3. Ask questions as a customer would.
4. Test common customer requests.
5. Test questions related to the agent's goal.
6. Test situations where a human handoff should occur.
7. Review the responses.
8. Improve the agent's knowledge when answers need correction or additional context.

Understanding the feature

Best practices

1. Give the agent a clear goal
   Define what you want the agent to accomplish rather than simply asking it to "help customers."

2. Keep knowledge accurate
   Add current business information so the agent can provide useful answers.

3. Define clear handoff rules
   Identify situations where a human should take over.

4. Test realistic conversations
   Ask the kinds of questions your customers actually ask, including difficult or ambiguous ones.

5. Start with approvals enabled
   Review AI actions during the initial rollout before allowing more autonomous behavior.

6. Choose channels carefully
   Activate the agent only on channels where its behavior has been tested.

7. Review the agent regularly
   Update its knowledge and configuration when your business information or customer workflow changes.

Common scenarios

Answering common questions

A customer asks about a product, service, or business process. The AI agent uses available knowledge to provide an answer.

Qualifying a lead

The agent asks relevant questions to understand the customer's requirements and qualification.

Booking an appointment

The agent guides a customer toward booking an appointment when that is part of its configured goal.

Handling after-hours enquiries

The agent responds to customers outside normal business hours when Hours is configured accordingly.

Human handoff

A customer asks to speak with a person or the conversation requires human attention. The agent hands the conversation to a team member according to the configured handoff rules.

Common mistakes

Giving the agent an unclear goal

A vague objective can make conversations less focused. Define a specific outcome.

Relying on instructions instead of knowledge

Business information should be added to the agent's knowledge rather than repeatedly embedded in long instructions.

Not testing before activation

Always test common and difficult customer scenarios before relying on the agent in live conversations.

Missing handoff rules

Without clear handoff conditions, conversations that require human attention may not transition as expected.

Enabling too many channels immediately

Start with channels where the agent has been tested and expand gradually.

Frequently asked questions

Is an AI agent created automatically?

One AI agent is created during onboarding, which you can then configure from Agents.

What can an AI agent do?

An AI agent can reply to customers on connected channels, answer questions using your knowledge, qualify leads, and book appointments when configured for those tasks.

Can the AI agent reply in different languages?

The agent can reply in the customer's language when it can.

Can I choose where the AI agent is active?

Yes. Use Channels to choose the connected channels where the agent should operate.

Can I limit when the AI agent works?

Yes. Hours can be configured so the agent is always on or operates only outside business hours.

When should I use handoff rules?

Use handoff rules for topics or situations where a human should take over, such as when a customer requests a person or the AI is not confident about the answer.

Related concepts

• Review AI Approvals
• Add Knowledge
• Assign Conversations and Hand Off from AI
• Using the Unified Inbox
• Leads and Customers
• Lead Scoring
• Manage Your Pipeline
• Appointments
• Automations
• Connect WhatsApp
• Connect Instagram
• Connect a Facebook Page
• Connect Threads