Intents

An intent represents the user's purpose or goal behind their input. It's a fundamental concept that enables MEVA to understand and respond appropriately to user requests. Intents are identified through techniques that analyze the user's words and context to determine what action the user wants to perform. For example, in a travel booking system, common intents might include booking a flight, checking flight status, or canceling a reservation. By accurately recognizing intents, the AI can provide relevant responses and execute the necessary actions, making the interaction more intuitive and efficient for the user.

This process involves analyzing various elements, including entities, phrases, and confidence scores.

Entities represent specific pieces of information or parameters within user queries. MEVA identifies these entities by recognizing keywords or phrases that indicate relevant details such as product names, or user preferences. By extracting entities from user inputs, MEVA gains a deeper understanding of the context and intent behind each query.

Phrases play a crucial role in determining the user's intent by providing additional context or clues. In general, phrases are used to identify the flows. MEVA analyzes the structure and language of user inputs to identify recurring phrases or patterns associated with particular intents. This analysis helps MEVA accurately classify user queries and generate appropriate responses. For example, to book a flight, we can add phrases such as,

"I want to book a flight to Paris for next week." 
"Can you help me find a flight to Tokyo?" 
"I need to reserve a seat on a flight to London." 
"What are my options for booking a flight to Sydney?"

Confidence scores measure the level of certainty or accuracy in MEVA's understanding of user intents. Higher confidence scores indicate a greater likelihood that MEVA has correctly identified the user's intent, enabling it to provide more reliable responses.

How Do Intents Work

The Intents section functions like a library, offering a structured collection of predefined intents that serve as foundational blocks for constructing conversational AI flows. This library provides a standardized framework of intents that can be utilized to design and develop MEVA interactions efficiently.

These predefined intents cover a variety of common scenarios such as greeting users, handling inquiries, processing transactions, and more. By leveraging this library, users can quickly construct conversational flows without starting from scratch, ensuring consistency and reducing time.

Additionally, users have the flexibility to add their own custom intents, allowing them to tailor the AI to specific needs and unique use cases. This approach not only streamlines the development process but also ensures that MEVA can adapt to diverse requirements, making it a versatile tool for various industries.

How To Add Intent in the Ephanti Application

To add an intent to the application,

1. Go to "Meva Assist-> Intents".
2. Click on the ellipses and select “Import”.
3. A pop-up window appears, click on the template link.
4. Add intents in the excel file and download as .csv file.
5. After downloading the .csv file, upload the file in the import pop-up.
6. You can drag and drop the .csv file or, you can upload it by clicking on the “Upload file” link.
7. Click on Next.
8. Here, click on “Validate&Import”.

The intent file is now imported and can be viewed in the Intents page.

"Intents" have been successfully imported which can be used while creating the chat flows.