IBM Watson at work

There may be much hype around artificial intelligence but IBM has been demonstrating the business potential of machine learning through its Watson analytics platform since 2011.

Watson came about after one of IBM’s research teams were at the pub watching “Jeopardy!” on TV and had the idea to create a programme that could compete with some of the grand masters of the quiz show game. It took nine years to build a comprehensive artificial intelligence solution that was capable of reading the quiz board, following the management of the show by the compere, and answering the questions. However, while it may have taken nearly a decade to create IBM’s Jeopardy solution it not only worked, it beat two of the grand masters of Jeopardy! in a live TV show that is now part of US pop culture.

IBM’s jeopardy solution, named after the founder of IBM, Thomas J Watson, received significant amounts of press but tasking sales forces to go out and sell it was not an easy task at first. However, IBM persevered and worked with companies on various successful projects, including some which only used parts of the Watson solution.

Today there are numerous examples of Watson at work. The Weather Company, one of the most used app on phones in the US, is constantly collecting data from “internet of things” sensors on planes, ships, cars and other devices from all over the world. This data is pooled in fractions of a second and questioned by end users billions of times a day with help from Watson. Twitter has “infused” some of Watson’s APIs into its own software to read and check Tweets for cyber bulling. While Vodafone is using Watson to help create TOBi, a “chatbot”, that can handle digital chat and answer questions with the telecommunications group’s customers in natural language.

While Woodside, an Australian oil and gas company, has used Watson to build an application that comprises tens of thousands of documents and manuals linked to the building and operation of oil platforms that can be interrogated in real time. Do you need to know if a helicopter carrying a certain weight can land on a rig when the wind is blowing at a certain speed? Normally such a question could require the expertise of multiple people, with different skill sets, but with Watson an operative at Woodside can ask that question and get an answer in real time. These types of applications of the technology could have immense use for companies that have a significant compliance burden and could benefit from a platform that allowed “natural language” interrogation of thousands of regulatory documents.

When finding practical applications for business, IBM Watson has determined the best method is to work together with the company to find solution and then start small and then scale up where possible. Woodside started with only one Watson-based application and now has several.

Volume and scale issues aside, however, there is almost no limit to Watson’s potential to provide a solution to a business problem in any sector. In agriculture, a sector which might appear at first to have little need for artificial intelligence, could benefit from drones trained to identify and treat particular diseases in specific areas of crops for entire farming regions. IBM Watson is heavily involved in the automotive sector and is working with Daimler AG to install chat capabilities in its cars and so a driver can get immediate responses to questions in natural language about their cars.

With so much potential, companies owning data that needs to be protected, do not need to be wary of using IBM Watson as the company’s approach allows users to have complete control of their data, provides separate platforms in the cloud to protect their commercial advantage and has all the contracts needed to keep legal departments happy.

It is clear Watson’s development so far has benefitted from a “perfect storm” where an explosion of data has coincided with significant advances in natural language processing, statistical machine learning and super-fast computers – and this rapid evolution is only going to continue.

Watch the video of Craig Dawson’s fireside chat: