A 6-Step guide to creating your first agent with Dialogflow in minutes

1. Ideation

Coming up with an idea before creating an agent in Dialogflow is a crucial step to ensure that your chatbot or AI system is built quickly, meets user needs, and operates efficiently.

Each agent is designed to solve one or more specific problems. For example, an agent for banking services will have completely different objectives compared to an agent for educational support.

Therefore, you need to define the goal, target users, and necessary features of the agent. Otherwise, you may lose direction, waste time, or create an agent that does not meet real-world requirements.

In this example, we will create a basic agent that responds to the user's question, "Hello, what is your name?"

2. Creating an Agent

Once you have an idea, proceed to create an agent in Dialogflow by following these steps:

Visit https://dialogflow.cloud.google.com/ and sign up using your Google account.

If you already have an account, you can skip the registration step and log in directly.

Accept the terms and conditions and continue.

After successful registration, you will see the following screen.

Click on "Create Agent," and you will see the following screen.

Fill in the required information for your agent:

• • Agent Name: The name of the agent you want to create. For example, "MyTestApp" or any name you prefer.
  • Default Language: Choose the language for your agent (e.g., Vietnamese or English).
  • Default Time Zone: Select the appropriate time zone. In this case, we choose Asia/Bangkok since Vietnam is in GMT+7.
  • Google Project: Create a new Google Cloud Platform project or use an existing one.

The last option is "Set as Mega Agent," which allows you to combine multiple agents (sub-agents) into one (Mega Agent). This is useful for managing different modules within a large chatbot system. However, if you are a beginner, you can skip this option for now.

After filling in the details, click "Create" to successfully create your first agent!

3. Creating Intents

Once the new agent is created, Dialogflow will direct you to the Intent creation page.

You can skip this step, but doing so means your agent is only 50% complete. It would be like baking a cake without filling, so let’s go through the process of creating an intent.

Simply put, an intent represents the objectives that you want the agent to recognize when a user interacts with it.

Examples:

1. User says: "Hello" / "Hi there" / "Good morning" / "How’s it going?"
   • Possible intent: Greeting
2. User says: "What’s the weather like today?" / "Will it rain tomorrow?" / "Show me the weather forecast."
   • Possible intent: Check Weather
3. User says: "I want to book an appointment." / "Can I schedule a meeting?" / "Arrange a consultation for me."
   • Possible intent: Schedule Appointment

To create a new intent, click "Create Intent." You will see the following screen.

In the Intent Name field, enter a name for your intent, such as "Greetings" for greeting-related messages or any name you prefer.

Click Save to save the intent for further customization.

4. Creating Training Phrases

In the Intent page, you will see multiple sections. For now, focus on Training Phrases and Responses.

Each intent in Dialogflow consists of training phrases (examples of what users might say) and responses (the agent’s replies).

In the Training Phrases section, click "Add user expression" and enter sample user inputs (e.g., "Hi", "Hello", "Good morning").

For this example, add: "Hello, what is your name?"

Click Save to save your input.

5. Creating Responses

Scroll down to the Responses section and find Text Response to add the agent's reply when it detects a user’s question.

For this example, add: "Hello! I am a bot created by BENOCODE."

Click Save to confirm your changes.

Now, whenever you ask, "Hello, what is your name?" the chatbot will respond, "Hello! I am a bot created by BENOCODE." Simple, right?

6. Testing the Agent

Finally, you can test your agent by entering "Hello, what is your name?" in the "Try it now" field next to the Save button.

If the response "Hello! I am a bot created by BENOCODE" appears, it means your agent is working correctly.

If the answer does not appear or an error occurs, you might have made a mistake in one of the steps. Try checking and correcting your setup.

If you have more time, you can add more Training Phrases and Responses accordingly. Also, don't forget to read more articles from BENOCODE to better understand how to use Dialogflow!