Bella’s technology allows your customers to lead the conversation with natural, human-like interactions. She can answer up to 60% of the calls for the front desk team, freeing them to interact with on-site guests and saving on fixed labor expense. Bella also offers an accurate understanding of your guests’ most common needs through her Precise Guest Insight™ tool.

Bella is based on a new and exciting approach to Natural Language Processing, a method known as ConveRT.  The system most of us use every day when we speak to Siri, Alexa, and Google Assistant relies on a different model that answers queries by quoting Wikipedia and other sources. This approach is not ideal for real-world conversational AI applications because it is computationally heavy, slow, and the required training is expensive.  We’ve used an entirely different and much more effective method to power Bella.  Bella is trained using social media conversational data for training, an ideal framework for this application. It’s effective, affordable, and quick to train. ConveRT achieves state-of-the-art performance across widely established response selection tasks, plus the more it’s used, the stronger and more accurate it becomes.