AI Agents
Overview
Section titled “Overview”AI Agents in Wylto are intelligent assistants that automatically handle customer conversations on WhatsApp. They can answer questions, qualify leads, and provide support — all without human intervention.
Creating Your First AI Agent
Section titled “Creating Your First AI Agent”-
Click the AI Agents tab from the main dashboard. You’re redirected to a guided AI Agent creation screen.
Navigate to AI Agents
Section titled “Navigate to AI Agents” -
Enter your Business Name and select the type of AI Agent you want to create (Q&A, Sales, or Customer Support).
Enter Details
Section titled “Enter Details” -
Click Create AI Agent to finish.
Create AI Agent
Section titled “Create AI Agent”
Creating Another AI Agent
Section titled “Creating Another AI Agent”You can create multiple AI agents to cater to various roles.
-
Click the existing AI Agent name displayed at the top.
Open the Agent Switcher
Section titled “Open the Agent Switcher” -
From the drop-down menu, select Create New AI Agent.
Create New AI Agent
Section titled “Create New AI Agent” -
Enter the Agent Name, select the AI Agent type, and click Create.
Name and Type
Section titled “Name and Type”
Adding AI Credits
Section titled “Adding AI Credits”Each AI Agent response consumes a certain number of AI credits based on the selected AI model.
-
Click the Credits section in the top-right corner of the dashboard.
Open Credits
Section titled “Open Credits” -
View the available credit packages and select the one you need.
Choose a Package
Section titled “Choose a Package” -
Complete the transaction to add credits.
Complete the Transaction
Section titled “Complete the Transaction”
Use this section to also view your AI credit balance and transaction history.
Setting Up an AI Agent
Section titled “Setting Up an AI Agent”To set up and deploy an AI Agent, you configure its settings, add a knowledge base and train it, define actions, test it in the Playground, and connect it to a channel.
Configure Agent Settings
Section titled “Configure Agent Settings”In AI Agents, click the Settings tab and open General. Configure the fields as required:
- Agent Type — Select the type of AI Agent: Q&A, Sales, or Service.
- Business Name — Enter the business name associated with the AI Agent.
- Include Business Info in Agent’s Knowledge — Toggle to decide whether business info should be used while answering queries.
- Support Multiple Languages — Toggle to decide whether the agent responds only in English or in the user’s language.
- Greeting Message — The initial message the AI Agent sends when a conversation starts or the user sends a generic message.
- Instructions to Agent — Define how the AI Agent should respond and behave.
- Custom System Prompt — Enable to override all the above settings and define a custom set of instructions.
Click Save to apply the General settings. Then:
- Open the Business Info tab, enter the required business details, and click Save.
- Open the Model tab, select the GPT model as per your requirement, set the Temperature, and click Save.
Add a Knowledge Base and Train
Section titled “Add a Knowledge Base and Train”The Knowledge Base is where you add the information the AI Agent uses to answer queries.
-
In AI Agents, click the Knowledge Base tab.
Open Knowledge Base
Section titled “Open Knowledge Base” -
Add content using any of the following:
Add Content
Section titled “Add Content”- Enter text manually
- Upload a document (PDF or file)
- Import content using a website URL
-
The added content appears in the section below. Click Train Agent to train the AI Agent.
Train the Agent
Section titled “Train the Agent”
Define Actions
Section titled “Define Actions”Actions define the tasks the AI Agent performs during a conversation, such as triggering workflows, sending HTTP requests, or updating contact details.
-
In AI Agents, go to the Actions tab and click Add Action.
Open Actions
Section titled “Open Actions” -
Select an Action Type
Section titled “Select an Action Type”- HTTP Request — Send or receive data from an external system.
- Trigger Workflow — Start a predefined workflow.
- Text Response — Give a text response for the agent to work upon.
- Add Tag to Contact — Add a tag to the contact during the conversation.
- Update Contact Status — Update the status of the contact.
- Update Contact Group — Update the group of the contact.
- Configure the action based on the selected type, then click Save Action.
Test in the Playground
Section titled “Test in the Playground”-
Go to the Playground tab in AI Agents.
Open Playground
Section titled “Open Playground” -
Start chatting with the AI Agent as you normally would.
Chat with the Agent
Section titled “Chat with the Agent” -
When an action is triggered during the conversation, click the action within the chat window to view the action logs.
View Action Logs
Section titled “View Action Logs”
Deploy the AI Agent
Section titled “Deploy the AI Agent”-
Go to the Channels tab.
Open Channels
Section titled “Open Channels” -
Select the desired channel (currently only WhatsApp is available) and click Connect.
Select a Channel
Section titled “Select a Channel” -
The available WhatsApp connections in Wylto are displayed. Click Connect WhatsApp.
Connect WhatsApp
Section titled “Connect WhatsApp”
Agent Types
Section titled “Agent Types”The Q&A Agent is designed to answer customer questions using your knowledge base.
Best for:
- Answering FAQs
- Sharing business information (hours, location, policies)
- Product information lookup
How it works:
- Upload documents, PDFs, or website URLs as your knowledge base
- The AI processes and indexes your content
- When a customer asks a question, the agent searches the knowledge base and responds with relevant information
The Sales Agent helps qualify leads and drive conversions.
Best for:
- Lead qualification
- Product recommendations
- Demo booking
- Price inquiries
How it works:
- Configure qualification criteria and product catalog
- The agent engages prospects with relevant questions
- Qualified leads are routed to your sales team or booked automatically
The Support Agent handles customer support conversations.
Best for:
- Troubleshooting common issues
- Order status inquiries
- Refund and return requests
- Escalation to human agents
How it works:
- Define common support scenarios and resolution paths
- The agent handles routine queries automatically
- Complex issues are escalated to human agents with full context
AI Credits
Section titled “AI Credits”AI Agents consume AI Credits based on usage:
- Each AI response uses a certain number of credits depending on the complexity
- Credits are included in your subscription plan
- Monitor your credit usage from the Dashboard
- Upgrade your plan if you need more credits
Best Practices
Section titled “Best Practices”- Keep knowledge bases updated — Regularly update your documents so the agent has the latest information
- Set clear fallback behavior — Define what happens when the agent cannot answer (e.g., transfer to human, collect contact info)
- Monitor conversations — Review agent conversations in the Inbox to identify areas for improvement
- Use specific knowledge bases — Focused, well-organized content leads to better AI responses