Earlier this year, Tridant’s Associate Consultant, Raymond Chan was invited to a 3-day training program with IBM in Melbourne, where he learned about Watson Explorer, a tool that enables users across an enterprise to search, navigate, uncover and share data-driven insights.
Watson Explorer can be defined as an enterprise dedicated search engine that combines data from internal and external sources, accepting both structured data (databases, data warehouses) and unstructured data (email archives, social network posts, web pages, file systems, etc.), and presents the information to users in a single view.
It is primarily used in enterprise-wide exploration, where users have rapid access to a wide range of data to discover insights, and as 360-degree information applications targeted for specific roles, and focus on key entities such as customers.
Watson Explorer Engine
The first two days of the training program were dedicated to the Explorer Engine – the original component of Watson Explorer, which serves as a web-based enterprise search platform with an interface similar to the Google search engine. The Engine is where data is extracted from defined sources, before being converted into a suitable format for the indexer to serve search results.
Watson Explorer provides more functions than just a web search
Clustering automatically organises search results into groups of related content, allowing users to have an overview of trending topics, as well as drilling down to low-ranked but valuable results.
A feature that I found particularly useful is the ability to collaborate with other users to make search results more valuable. The Collaborative Search allows users to add annotations such as comments, rating or tags, which can result in a considerable amount of time saved for colleagues and future searches.
Naturally, search queries can also be saved, shared, and exported for use outside of Explorer.
360-Degree Information Applications
Another important Watson Explorer module is the App Builder, which was the main focus of the last training day. It connects to the users data from the Engine and displays it in a custom web interface, featuring 360-degree views of relevant entities such as customers, products, or locations.
Building an application appears to be easier than building the Engine, thanks to the less complicated user interface:
Relationships between entities are defined in a way similar to defining tables’ relationship in a database, and pages are created via drag and drop of widgets on the chosen layout. (Of course, some database and programming skills are still needed.)
The end-user’s interface can be customised with the enterprise’s colour and logo, and features a home page, search page, as well as pages for every entity. For example, searching for a particular customer will bring the user to a page containing the customer’s contact details, purchased products, recent conversations, and more.
One of the most impressive aspects of Watson Explorer resides in its ability to analyse and extract insights from natural language via a range of sources, including emails, comments or Tweets. Some widgets can be used as useful alternatives to functions in SPSS Modeler, and predicts the propensity of a particular customer to churn, or can gauge the customer’s current satisfaction toward a product or a company.
Applications can furthermore be enhanced with different API services from the Watson Developer Cloud which is regularly updated with new functions.
If employees started using Watson Explorer collaboratively, the potential will be there for it to become an indispensable tool for fast and efficient research, discovering insights from structured and unstructured data, and better decision-making overall.