Here is your Monday briefing:
New York City warehouse is facing a shortage of product A and product D. There is a shopping season coming in 2 weeks. We might see a huge spike in demand for product A and D.
Solution: Move the Product A from Chicago Warehouse and Product D from Boston Warehouse to New York Warehouse. Let me know if you want to initiate the process. (Yes or No)
There is a problem in the Manufacturing plant at Houston that the machine X is not working properly and it needs replacement.
Solution: Please release the order for the part X from either of these below vendors :
Vendor 1: Low Price, Medium Quality, Faster Delivery
Vendor 2: High Price, Highest Quality, Delayed Delivery
Based on the available data, I see that Vendor 1 suits our needs. Do you want to go forward and order the part?
The overall production of all the products is comparatively below than the Q1 expectations. If it continues, we will not meet the projected March demand.
Solution: Please send out a mail to the factories that are not aligned their production with forecasted demand.
Factory 1: by 10%
Factory 2: by 5%
The Enterprise Chatbot helps the director in many important business decisions such as maintenance of plants, supply-demand gaps, meeting the projections etc. I see that we all want such type of BOTs in our enterprise to increase our efficiencies.
Current Scenario
Before we even question when can we have this, Let us look at what are the things that are required to build this Enterprise BOT.
- Central Data System: Connects to all the different data systems and collects the data from those systems.
- New Business Rules: Helps to transform, and process the data. It varies across each department.
- Advanced Analytical Engine: Contains Machine Learning Models to churn the data inside the central database based on business rules.
- Powerful Natural Language Models: To understand complex queries and mood of the human nature to converse like a human.
- Generative Business Language Engine: To generate human type conversations on the fly based on the processed data.
Central Data System
Here the system collects and stores the data from different departments of an organization irrespective of whether the data is structured or unstructured. In addition to that, the system will clean, transform and load the data for different engines to use. I will not dig deep into the central data system as itself is a vast topic to cover. A few of the chatbot solutions that are hard for the organizations to crack are:
- Integration of the systems across the company to a central data system. In most of the big organizations, the departments work mostly in silos using their own data systems.
- Data migration from one department to another department is very limited. The major problem is due to the notion that the data from one department may not be of any use to the other departments. For Example, the ordering of a new system in a manufacturing plant may not be available to the product team. If we look at the same aspect in a different light, the product team may help to build and customize the system in a price effective way.
- Usage of legacy systems which may not integrate well with the present systems. For Example, some of the big companies in the banking industry are still using legacy systems for the transactions. The migration of the legacy system to the present technology itself will take more time than the integrations itself due to limited resources in that space.
- Different Formats of the data. Nowadays, the data is not only structured but also unstructured such as e-mails, images, videos, documents etc., This data is present not only in the enterprise systems sometimes it may reside in individual’s devices. Collection and Organizing the different formats of data in itself a big challenge.
Despite the challenges above, the organizations are realizing the need for this sort of system due to changing rules in the business landscape and want to be ahead in the adoption of the technology.
New Business Rules
Though the organizations are transforming, the real challenge is getting the insights across different departments from the central data system. In other words, what are the business rules that help us provide cross-departmental insights?
This is still a bigger problem to address as each industry contains a different set of standards, policies, and guidelines to follow. For the first time, people are looking at this sort of data from different perspectives. In addition to this, there are other challenges to lay down the rules i.e. What data needs to be accessed by other departments?
To overcome this, look at the present business rules and gather knowledge from other departments’ experts about
- How that business rule will impact their department?
- When will that business rule apply?
- Where in the business process that the business role comes into the picture?
- What is the information or parameters required to execute the business rule?
- Will it need any more changes to fit the other departments?
Answering the above will make us come closer to have a cross-departmental perspective of the business rules that will impact our organization. This may also help us to pave way for new business rules in the organization.
Advanced Analytical Engine
After the business rules, the next thing to look at is the Analytical Engine. What exactly is an analytical engine? It contains the Machine Learning Models to churn the data inside the central database based on business rules to provide valuable results that will impact the business.
Similar to the business rules, the analytical engine also needs to come up with the new KPI’s for different departments. The biggest challenge in the analytical engine is to collate the structured data and unstructured data based on new business rules. As the data now comes in many formats, processing, and understanding that requires different analytical systems
- Structured Data Analytical System
- Unstructured Data Analytical System
- Text Analytics (Mail and Documents)
- Image Analytics
- Video Analytics
Now, based on the above systems, Analytical engine helps us to derive the following analytics efficiently.
- Descriptive Analytics: It helps us to know what is happening in the organization. For Example, a sales manager will know about the sales for the present month. It is primarily focused on what is going on and whether it is going in the right or wrong direction.
- Diagnostic Analytics: It compares the present data to the previous data to know why something happened in the organization. This type of analytics is very important to know the underlying dependencies and patterns which leads to the present results. It will also help us to define the problems in the organization.
- Predictive Analytics: It will predict what will likely to happen in the future. It takes into account descriptive analytics and diagnostic analytics of the organization to forecast the trends. This sort of analytics helps the organization to plan for the future and avert any problems beforehand.
- Prescriptive Analytics: It will help to provide the solutions for future problems or to take full advantage of the trends in the industry. This type of analytics requires the use of external information other than the organization. It is very complicated to implement as it has to take a lot of information not only based on business rules but also on external factors.
Till now we looked at the back-end that powers the Future of Enterprise Chatbot, in the next part the future of enterprise bots, we will look at the front-end of the BOT which will help us to communicate and interact more like a human.
