Updated: Feb 9, 2019
Datalog.ai’s brav/ai™ platform includes a natural language understanding engine built within an easy to use tool for chatbot builders and developers of virtual assistants. It makes machine to human interactions more intelligent, personalized and conversational.
Today’s AI tools only respond to a few phrases or questions, and cannot maintain the flow of communication from beginning to end. To address this, some chatbot software developers will use temporary memory to capture words that might be helpful later in a conversation. This does not readily handle a large amount of simultaneous human interactions. This would be the case with a popular chatbot that may have millions of interactions per day. The current remedy to this is a cutover to real humans during a session, filling the gaps in other AI platforms.
This approach defeats the purpose of using intelligent machines at all. ~ Jack Crawford, CEO, Datalog.ai
The platform uses a novel and proprietary method to maintain the flow of a conversation from beginning to end. This “state” is maintained even if the conversation diverts temporarily to another topic. It remembers what you have said earlier and brings that information to future exchanges in a machine to human chat session. This eliminates the need to save an entire conversation as it progresses.
Another challenge is present in business. Enterprises have tremendous amounts of useful data locked away in various systems for logistics, finance and customer management. By necessity, corporations must diligently protect both customer information and business data from outside threats. Cloud-based artificial intelligence models require that this data be sent off to their systems, processed, and then returned with the response that is delivered from the machine to a human. Organizations are likely to resist sharing such data where it will be logged and analyzed for future use by the cloud AI provider.
The brav/ai™ conversational component addresses and solves this concern. It is available for deployment completely within the corporate network of a business, allowing safe access to useful information present within their enterprise.
A final challenge for the creation of intelligent machines lies in the immense data required to train models that would be able to carry on a conversation. Unfortunately, this magnitude of data is available only to a few organizations that have amassed personalized data over the past decade. Most of us lack access to such data stores and need a way to create intelligent, personalized and conversational chatbots.
This is where it brings the most novel capability to the market. It is able to learn how to converse within a focused topic with a very small amount of training text.
This innovation affords a more natural interaction with the machines we are creating ...
... which allows for the creation of intelligent, personalized chatbots for consumers and businesses.
Now we can truly benefit from the “virtual assistance” that modern artificial intelligence promises to each of us.
Copyright (c) 2017, Emma Davila, All rights reserved