Create, train, and nurture the best virtual assitants. Kore.ai helps enterprises more effectively analyze and perfect their VA’s task fulfilment and NLP performance.
Why businesses need to regularly monitor, analyze and train their virtual assitants
Today’s consumers frequently prioritize customer experience and service quality over other factors like product or price. Virtual assistants can play a key role in meeting high consumer expectations for service, but a poorly designed or maintained bot can ruin the customer experience. It’s vital that organizations monitor overall virtual assistants performance, analyze identified failure scenarios, and periodically retrain their virtual assistants in order to provide high-quality service, improve brand perception, and convert more prospects into customers.
Kore.ai provides a comprehensive analyze and learn module that helps you fine-tune your virtual assistants, so you can improve the customer experience, more effectively engage new audiences, and efficiently automate sales, support, and other key functions.
Review, iterate, and perfect your virtual assistants with Kore.ai
Kore.ai enables you to perform root cause analysis on various failure scenarios so you can improve virtual assistant performance and accuracy via targeted bot learning. Our analyze and learn module provides all the tools and visualizations VA creators and business analysts need to better understand bot usage patterns, conversation flows, and drop-off points. Enterprises can even review and export chat transcripts and other detailed conversation data in order to better understand failure context.
Identify the right intents
Get in-depth insights into your VA’s performance in terms of identifying user intent, so you can quickly fix the following common issues:
- Out of scope intents: Occurs when users ask questions that your virtual assistants hasn’t been trained to answer.
- False negatives: Occurs when your virtual assistants already has a certain intent built but fails to recognize alternative phrasing or terminology.
- False positives: Occurs when your virtual assistant matches your intent incorrectly to another intent.
“User asks an IT help desk bot about the weather”
“Bot fails to match the question “how rich am I” to the VIew Account Balance intent”
“User wants to know how much mobile data they’ve used so far this month. They ask the bot to “get usage,” but the bot responds back with their total bill balance instead.
Improve task fulfillment
Monitor overall backend performance and view all tasks that failed to execute after matching with an intent, so you can quickly identify and fix the following common issues:
- Task failures or incomplete dialogs: Occurs when users abort a task, initiate an alternative task, or refresh a conversation window. Can also occur due to authorization failure.
- Service and script failures: Occurs when the platform is unable to make a proper API call to a backend system. Can also occur due to issues with a service provider, in which the platform is unable to provision a response JSON from the provider.
Create better, user-focused conversation flows
Map popular user paths and exit points with our powerful session flow visualization tool. This enables virtual assistants creators to view comprehensive details about messages and related tasks at a specific point in a conversation. Virtual assistants creators can even drill down to a specific user to view complete chat transcripts and debug context.