Skip to main content
We know that User Inputs are used to create flexible agents. However, instead of open text fields, you can present the user with a list of predefined options, ensuring that responses always stay within what’s expected. In this tutorial, we’ll explore how to use the Single and Multiple Selection fields to create a smart travel planning agent.

​
Understanding Selection Input Types

Single Selection Box (Dropdown)

Ideal for when you need the user to choose only one option from a fixed list. This ensures the response is singular and decisive (e.g., “What’s your budget? High, Medium, or Low”).

Multiple Option Box (Checkboxes)

Perfect for allowing the user to select multiple options from a list. It offers more flexibility, multiple preferences (e.g., “What are your interests? Food, Culture, Sports”).
Image
For both types, the list options are created the same way: just type them into the configuration field, separated by commas.

​
Our Example Project: The Travel Planner Agent

Let’s build an agent that works like a tour guide, creating a travel itinerary based on the user’s preferences.

Initial Agent Setup

In AI Studio, start creating a new agent. Since our agent will generate a text itinerary, a Chat or Text Agent is ideal. Keep the default “All LLM” selection or choose a specific model you prefer.
Image

Initial Agent Setup

In the prompt field, we’ll insert the logic for our tour guide, the context, and the baseline guidelines that will guide it:
Assume the persona of an experienced tour guide and provide the best ways to explore the city according to the type of trip and the activities chosen by your client. At the end, add a note saying how important it is to check local hours and rules before the trip. _City:
Trip Type:
Desired Activities: _ Critical: Create a highlighted theme for each chosen activity and offer more than one activity option. Do not mention the pandemic. Deliver only what was requested in the best way possible.
Image
The variables for city, trip type, and activities will be the connection points with our User Inputs.

Configuring User Inputs

Now, let’s create the interface our end user will fill out.
  • Input for City: “Short Text”, with the name city and the label “Which City and State do you want to visit?”
  • Input for Trip Type (Single Selection): “Single Selection” with the name trip type and the label “What is your trip style?”. For the options, enter the list of choices separated by commas. E.g.: Budget, Romantic, Adventure, Family, Luxury
Image
  • Input for Activities (Multiple Selection): “Multiple Selection” with the name activities and the label: “Which activities do you like the most?”. For the options, enter the list of interests separated by commas. E.g.: Culture and Museums, Food, Nightlife, Shopping, Parks and Nature
Image
Tip: Use the ”+” button in the prompt editor to insert the variables into the prompt and ensure an exact match by referencing the variable in training.
Image

Saving and Testing

With everything configured, name your agent and click “Save” and then “Preview” to test it. You’ll see a friendly interface with a text field, a dropdown menu, and checkboxes, ready to be filled out.
Image
This tutorial shows how selection inputs can make your agents more structured, guiding the user and ensuring you receive information the way you expect. Keep exploring and creating your agents!