Creating Your First TrustYou Agent

Welcome to our Quick-Start guide! This document walks you through setting up a minimal viable Agent (MVA) so you can experience TrustYou Agent AI in action. By following the steps below, you’ll learn the core concepts, create a simple Agent, add knowledge, and start chatting right away.

If you encounter any issue contact us.

Overview of Key Concepts

1. Agent
   • The AI chatbot entity that users interact with.
   • Has a name and a specific purpose or goal (e.g., booking assistant, guest service, etc.).
2. Places
   • Each Place can represent, for example, a hotel, a restaurant, or any other location/resource for which you want the Agent to have knowledge on and answer questions or give recommendations for.
   • An agent gets Places assigned they should know about and use that knowledge within the conversation to achieve their goal
3. Resources
   • The actual content and data (websites, documents, reviews, etc.) tied to a Place.
   • By assigning Resources to a Place, the Place can “feed” the Agent the knowledge it needs to answer user queries accurately.

So in Short, an Agent has a goal it tries to achieve by leveraging the knowledge they have on places they have been assigned to.

Quick Navigation of the Backoffice

1. Home
   • Provides an overview: recent activity, conversation summaries, and stats on how many Agents, Places, and Resources are configured.
   • Perfect place to see your current progress and usage at a glance.
2. Agents > Overview
   • View, create and manage your organizations Agents.
   • You can also import an existing Agent if you want to get started extremely quickly without going through all steps manually.
3. Agents > Insights
   • See usage statistics on you organizations Agents
4. Knowledge > Places
   • View, create and manage your Places
   • This is also where you attach knowledge to individual Places
5. Knowledge > Resources
   • Global knowledge resources not limited to a single Place

Step-by-Step Setup

Below is a concise process to create your first minimal Agent and get chatting. We recommend short tutorial videos or diagrams at key points to help illustrate each step.

1. Create a New Agent

1. Go to Agents > Overview in the navigation bar.
2. Click on Create New Agent.
3. Enter a name (e.g., “BookingAgentTest”) and select its purpose or goal by specifying its type.
4. Save your new Agent.

2. Create a Place

1. Navigate to Knowledge > Places.
2. Click on Create Place.
3. Give the Place a name that corresponds to its real-world counterpart (e.g., “Dorinth Hotel Köln”).
4. Select the appropriate Type (Hotel, Restaurant, etc.).
5. Save your Place.

3. Add a Resource

1. Select your newly created Place and click Create New Resources.
2. Enter a name (e.g., “Dorint Website Info”) and select the Type of content (e.g., Website, Documents, Reviews).
3. Paste in the link, upload a file, or connect to review data.
4. Choose the Extraction Pattern (e.g., “Hotel Facts,” “Raw Text,” or “Generic Knowledge”). This helps the system better parse the content and know what it is looking for.

4. Assign the Resource to the Place

If you created your Resource from a Place (as described above), you can skip this step. If a Resource has been created via the global Knowledge > Resources tab, you have to add it to you Agent as it has been created as a global Resource.

1. Select the Place you want to add the Resource to.
2. The global Resource should show up, simply select it and save the Place.

5. Link the Place to Your Agent

1. Go to Agents > Overview and select your Agent.
2. Click on Places.
3. Select your Place and save the Agent.

6. Preview & Test the Agent

1. Go to your Agent’s detail page and click Preview Chat .
2. Open the preview to start interacting with your Agent.

Best Practices & Tips

• Start Small: For your first test, use one Place and one Resource to keep it simple.
• Validate Accuracy: Check if your Agent can answer basic queries correctly. If not, adjust your Resource extraction settings or add additional content. You can even manually add Knowledge by adding a Local Resource
• Iterate: Once comfortable, you can import more complex Agents or add multiple Places and Resources for richer experiences.
  
  

Where to Go from Here

• Add More Resources: Link multiple documents, websites, or review sources to each Place for a robust knowledge base.
• Refine Extraction Patterns: Experiment with different extractors (e.g., “Hotel Facts,” “General Knowledge”) to see which yields the best answers.

Conclusion

You’ve now set up a minimal viable Agent, linked it to a Place, and enriched it with a Resource. With this foundational knowledge, you can experiment with more advanced features, refine your Agent’s purpose, or add more Places and Resources.

If at any point you get stuck, check the preview widget or revisit your extraction patterns. For an even faster setup, remember you can also import a pre-configured Agent to jumpstart your testing.

If you have any further questions or need deeper troubleshooting, be sure to check our Knowledge Base or contact us via Support.