Big data, machine learning and artificial intelligence are converging into cognitive platforms - natural language query engines that have the potential to transform entire industries and even professions. David Reed talked to Phil Westcott, Watson eco-system leader for Europe at IBM, about the impact its solution is starting to have.
DataIQ: IBM talks about an eco-system developing around Watson - what is that and how does it work?
Phil Westcott (PW): Companies are getting excited about our cognitive technology, so we have made it available to an eco-system of partners that can package it and drive it into different industries and markets. We have 400 developers worldwide, using our technology to solve problems in a wide range of different industries.
The eco-system is all about getting into new areas - one partner has built a pet adviser which faces directly to the consumer. That is not something which IBM would necessarily have developed and shows what effect our approach is having.
Quantone Music is another good example. Its MusicGeek solution uses Watson in the music industry to build musical taste into recommendations. They are going to large labels and offering them the ability to “bottle their taste” in music, providing what their customers will want to hear. It is not Watson that has the taste - it just provides those clients with the ability to pull it together out of their music assets.
DataIQ: So how do partners join that eco-system and how is it run by IBM?
PW: We are two years into Watson - it was launched on 9th January 2014 as a commercial entity - and so far we have hand-picked those partners. They are innovative players in their industries who come to see us. From there, it has taken on a life of its own with partners diving into the technology themselves and building “powered by Watson” applications.
We do guide them, keep them up-to-date and provide an over-arching strategy. We have broken out Watson’s services into 28 APIs which enable those partners to build their solutions on our platform. The key is to allow them to run with the opportunity and give us a global reach we couldn’t otherwise achieve.
DataIQ: How do you see cognitive platforms getting built into decision support?
PW: The way an expert makes a decision is that, when they are presented with a situation, they draw upon their domain expertise and experience as well as the context of that situation. In any discussion, they will be making a decision based on a broad background of knowledge and experience. So decision making is a combination of knowledge and real-time opportunities.
A cognitive platform can draw on data from a “back catalogue”, such as a static knowledge base, as well as real-time signals. Watson automates that expertise and knowledge by processing the available data, running predictions and presenting an optimal answer.
DataIQ: So how is that impacted by the fact that 90 per cent of information is in the form of unstructured data?
PW: Leveraging that 90 per cent is critical because that data is typically an asset for the company. You can plot it in a simple two by two matrix of real-time versus historical and internal versus external. Unstructured data can be very value-adding to the decision-making process because you can use it as reference material. There may be unstructured data that is critical to the organisation and which is internal, but it can then use Watson to pull in external unstructured data from the Web. In the past, because that data was unstructured, it was not possible to develop it as part of decision support.
DataIQ: What needs to happen in an organisations for it to take advantage of your cognitive platform?
PW: The opportunity is for an organisation which is already using a level of analytics to become a cognitive organisation. That can start in a number of ways - one of the first is getting a grip on unstructured data inside the organisation and making it visible and digestible to staff so they can use it. It is about applying cognitive resources to knowledge management, which is what the Watson Explorer platform does.
As the organisation starts to bring in cognitive APIs, it will begin to see patterns in that unstructured data. Then the key is to merge unstructured and structured data into business intelligence.
DataIQ: Is there a fear factor around making that change?
PW: It is a cultural change and a new era of computing. We have a high expectation that it will change the way people interact with technology.
DataIQ: Are you seeing an appetite among the new generation of managers for this type of technology?
PW: One of the reasons for our eco-system is to create solutions which can get managers started on that journey. It can come from the top down using an enterprise approach to drive change in the whole business, or it may get on the road by embracing one of the 100-plus solutions which are in market. If they do that, it will help with the cultural shift involved.
DataIQ: Early benefits are starting to be seen - when do you think cognitive platforms will really take off?
PW: It is always hard to gaze into the future, but the way we interact with cognitive systems will be very different. There are human elements that are unique and will always be critical, such as moral judgement or the ability to solve non-hierarchical problems. The blending of man and machine will bring about amazing changes over the next decade and beyond. It will enable us to process large quantities of structured and unstructured data to make better-informed decisions and understand markets in ways that have never been possible before.