It’s interesting that the semantics of language vastly differ according to a person’s position in an organization.
Most stakeholders have access to central information systems in an organization. They are given customized views of data within their roles and hierarchies. Often times, they end up asking the IT system the same question, but in a different way. Previously, business intelligence interfaces had multiple dashboards and reports for different roles ensure the system made sense.
A Centralized Solution
However, imagine if there was one person who could answer the questions of all these stakeholders. How would they end up communicating with him?
Now, imagine if the Head of Sales wants to know more about his team’s sales numbers for yesterday. He would probably ask, “What are my sales for yesterday?”
What if the CFO wants to know the answer to the same question? He might frame his question as, “What is the company’s revenue for yesterday?” Also, the truck driver might want to know what were the number of shipments he carried yesterday. He might ask, “What were my shipments for yesterday?”
As you can see, all these stakeholders are requesting the same information. Since only one person knows the role and hierarchy of each of the three individuals, he ends up sharing contextually relevant answers with each of them.
However, if this were an AI entity his answers would automatically go through a learning curve to make them contextually relevant.
How To Support The Conversational BI Learning Curve?
Conversational BI assistants use different adaptive learning techniques like supervised learning and reinforcement learning. In essence, this gives them the ability to perform multiple tasks.
Reinforcement learning is a trial and error approach. Then it’s logical to ask, what happens when the assistant is not able to understand the question? It tries to provide options on the current context to help the dialogue progress more effectively. Moreover, this process helps the assistant to deliver context to the dialogue. In the end, the user benefits from the customized information based on their specific needs.
Supervised learning teaches AI assistants by feeding them with training data and clarifying options with SME knowledge. Therefore, the assistant can understand and answer the end user next time. By this way of training, the assistant gets smarter and improves its answer accuracy.
For example, customer service departments are using AI assistants to answer their end customers’ questions. The AI assistants are trained by making them scan through thousands of hours of customer-agent call records. Ultimately, this helps build context and humanize the conversation.
Conversational BI platfroms like ConverSight.ai from ThickStat use a combination of these learning methodologies. The result is powerful intuitive interfaces that serve their customers’ information needs. If you want to learn more, take a look at the recording of their recent webinar here – https://www.youtube.com/watch?v=l64f5vreKCc