Virtual agents continuous learning – What makes the user experience truly positive? Response speed, convenience, consistency, friendliness, but also empathy and “human touch”.

These are the elements that emerge from PriceWaterhouseCooper’s “Future of CX” survey, conducted on a sample of 15,000 people in 12 countries around the world (including 4,000 within the United States).

The research shows that 73% of people interviewed indicate the customer experience as the most important element in the orientation of their purchasing decisions.

The data collected also shows that consumers appreciate suppliers, brands and products that offer a user experience with no inconvenience or frustration, where efficiency and promptness of response times are made possible by the new AI technologies.

People expect the technology to be functional and user friendly while automation is expected to facilitate interaction and to provide instant response times. We know that virtual agents are able to provide immediate answers to customer requests and interact in natural language, however the main element that discriminate between a good and a bad user experience is the understanding capability.

Virtual agents customer experience: “continuous learning” functionality.

Virtual agents continuous learning – Virtual agents learn automatically by analyzing data from past interactions and formulate hypotheses about the right answer. However, if it is true that making mistakes helps learning, bots have to make many mistakes to learn.

While virtual agents are developed with a clear image in mind of the customer journey, once in production the bots have to manage all kinds of user utterances and it is important to speed up learning with less mistakes possible to provide a good user experience.

For this reason, the concept of “human in the loop” involves human intervention in a “supervised” learning model, in which the bot’s responses are controlled and modified by humans, allowing the bot to learn fast with less mistakes.

To further facilitate and speed up the learning of virtual agents, CRAFTER.AI™ platform integrates the “continuous learning” function, thanks to which “misses” messages are automatically reviewed, giving the user the possibility to ignore them or add them automatically to the bot’s knowledge base, in the form of a new intent or integrating an existing one.

In this way, CRAFTER.AI™ users become even more independent in updating their virtual agent.

The most advanced technologies do not lead to any results if they do not take into account the needs of those who use them.

In particular, when talking about conversational AI, virtual agents understanding capability is a fundamental element for the customer experience, just as the ability to monitor and update the knowledge base in an easy and independent way is a precious requirement to improve the user experience of our customers.


You may like: