Today’s world is built on information. Everyday data is generated or collected, with information on facts, figures and statistics continuing to grow. With so much data available, it becomes increasingly hard to know which data has value, and where exactly to look. The universal problem experienced by companies of all sizes is access to vast amounts of useful customer and market data masked (or engulfed) with valuable insights. There’s simply too much information to process.
A global health-technology company is facing the same problem, who over a three-year period relaunched over 110 of their customer-facing websites, each of which featured performance measuring tools and dashboards. With all this data of traffic, user activity, and demographics, the average user doesn’t know what to look for. It becomes very difficult to try to filter down and examine it, and the data never lives up to its full potential.
With new innovations in the space of BOT, SmartBots have overcome this challenge — by giving BOT access to information, establishing a simple way for users to interact with it, and providing a platform from which it can present its findings clearly and concisely, this can unlock a true treasure trove of information. It sounds easy on the surface, but it’s actually complex. Firstly you need a BOT that’s simple to use, intuitive, and human-like. Secondly, it needs to be able to understand the context, learn on the go, be able to engage with users, and finally present intuitive responses. Significantly more advanced than their predecessors, SmartBots are built with four core features at their foundation; Smart Intent Orchestrating, Smart Context Handling, Smart Response Handling, and Smart Learning.
- Smart Intent Orchestrating enables the BOTs to be proactive, rather than merely reactive. When queried by a user, the BOTs can engage in conversation, help naturally progress a dialogue, and also try to provide guidance. This is achieved by recognizing user behavior and interest, understanding the correlation of information and offering suggestions based on this conclusion.
- Smart Context Handling allows BOTs to retain the ‘journey’ of a conversation with a user. Unlike conventional rigid BOTS, a SmartBot understands the expectation of results based on past context when the users drill further down into a dataset. As such, the SmartBots presents its findings in context with the dialogue flow.
- Simplicity is the key to Smart Response Handling. Recognizing that the answers provided would have little value if they could not be easily presented, a SmartBot refers to a knowledge base which directs the response in order and form to present information.
- Finally, Smart Learning means that a SmartBot has the ability to learn from each user interaction. Remembering the pattern of queries and replies, the SmartBot is able to recognize future users need for additional information. SmartBots pays attention to not just what’s being said, but also to what’s being NOT said, adjusting its behavior accordingly.
The ‘Smart’ features are just the fundamentals of the SmartBots. To really make outstanding BOTs, we built a framework which will make the BOT human-like. By utilizing a number of key Amazon Web Services, such as LEX, AWS Sagemaker, DynamoDB and Comprehend, a SmartBot not only satisfies all of its ‘Smart’ criteria, it advances to the next level. It is voice-enabled, capable of adding its own expression and personality into conversations, can ensure its speech is varied and responsive, and can even analyze the mood of the person interacting with it and tailor its responses accordingly.
As a result of the new innovations in BOT development, the SMART BOT is already in operation and having considerable success. It is able to access all the information gathered by the company’s various websites, analyze it, and clearly and efficiently present her findings. She also provides her own recommendations and advice to users, helping them delve deeper and get further insights. With SmartBots, the information the company gathers is simple and easy to access. Trying to get the same data without her, using original dashboards, would take three times longer. It would also be limited by user own queries, whereas SmartBot is able to offer intuitive suggestions and insights based on her own observations.
Perhaps a SmartBots best feature is that it is API (application programming interface) driven; with the right API in place, it can connect to anything. Then all users need to do is ask!
SmartBots demo at AWS re: Invent 2019 — https://youtu.be/6Fwoe6uG_Aw
For more details, visit www.smartbots.ai