By David Stokes
In 2013 it was estimated that 90% of all of the data in the world had been generated in the preceding two years. The vast majority of digital information (about 80%) is generated in an unstructured form – articles, blogs, posts on social networks and the like. In many ways, our digital world has simply become too human. Coping with this therefore requires a new approach – a system that can imitate the rational processes of a person, at the speed and scale of a computer. Enter IBM Watson.
We have become accustomed to watching technology change before our eyes, so it takes something truly revolutionary to make us sit up and take note. In February 2011 IBM Watson defeated two previous champions on the US game show Jeopardy! – the country’s leading general knowledge quiz show – and in doing so demonstrated to the world that a new kind of computing system was emerging, a system capable of transforming the partnership between humans and technology.
The subsequent years have seen Watson technology step off the game show floor, and into offices, hospitals and governments, across 17 industries and six continents. Watson supports systems that are not only attuned to, but also actively collaborative with the user, able to both direct and amplify their expertise. This marks a conceptual shift in the way we see computing: from a purely enabling force, to one capable of creating and contributing value independently.
The resources available to an organisation can be considered in many different ways: as both tangible assets like oil or skilled labour, and the intangible such as time and productivity. But one of today’s most plentiful and valuable resources is also one of the most frequently neglected: the huge volumes of data generated every second. Digital engagements are happening constantly, and the body of data they generate contains many of our most brilliant discoveries, most valuable observations, and most intriguing opinions. But if such a digital goldmine exists, why aren’t we doing more about it?[ms-protect-content id=”9932″]
Data on this scale is relatively new – in 2013 it was estimated that 90% of all of the data in the world had been generated in the preceding two years. As with many other technological revolutions, it can take time for us to realise how these innovations could impact our world for the better. Changes of this kind present challenges and disruptions, but these are part of the journey through which disruptive forces become the norm. In 1878 the success of the telephone brought the communications industry to a tipping point, and decimated the powerful monopolies controlling the telegraph industry. We believe that Watson hastens us to another such tipping point; changing the way we use and gain insights from Big Data.
Traditional computing technology can be applied with a degree of success to structured data – neatly organised input that systems have been programmed to process. However, the vast majority of digital information (about 80%) is generated in an unstructured form – articles, blogs, posts on social networks and the like. Imposing the inflexibility of a traditional computing system on this input is pointless; it is too nuanced, too irregular, too rooted in the context of a time and place. In many ways, our digital world has simply become too human.
Coping with this therefore requires a new approach – a system that can imitate the rational processes of a person, at the speed and scale of a computer. Enter IBM Watson.
Watson responds to unstructured natural language like a human, starting by analysing text with reference to context and semantics (rather than simply identifying keywords like traditional speech recognition software). This allows Watson to understand the intent of text, and direct further enquiries accordingly. Considering the body of its existing knowledge and its previous interactions, Watson evaluates then proposes the answers most likely to be correct. In this way Watson approaches a problem in the same way as a person: observing and interpreting the issue presented to it, and deciding on the correct solution with reference to its own knowledge and experience.
When introduced to an industry, Watson learns the “language” – medical jargon or the terminology of the insurance industry for example. Human experts then train Watson using the available literature, running through basic questions, answers, and relationships, to teach Watson how to approach problems. As the learning process continues, Watson’s progress is reviewed, and the accuracy of its solutions checked and refined. This dynamic learning process is key, and allows the system to continuously evolve and improve – while conventional technology eventually becomes outdated, Watson actually grows in value over time.
These capabilities form an enviable skill-set, but how is Watson being employed in the real world?
Some degree of internal division is necessary in every organisation – by departments, fields of expertise, or major accounts, in everything from governments to multinational corporations. This also produces divisions in their data, which is stored on different systems or in different formats. Producing an integrated perspective under these conditions is a vastly inefficient process, and individuals at all levels are left under-prepared for key decisions in areas such as strategy, efficiency, risk, and relationship management.
Watson is able to interpret the information from these separate internal systems and external sources to present a single, 360-degree view of an organisation. This is a system that can produce a perspective that even the most well resourced levels of a company might have previously struggled to put together, and provide it to any employee – from directors to front-line staff. This view can then be combined with analytic technologies to produce far more useful insights, insights informed by all areas of the enterprise and that consider the reactions of the unit as a whole.
Watson-enabled analytics do not attempt to impose a more comprehensive or uniform structure on an operation, but instead respond to the characteristics of each unique operating model – and will continue to respond to changes as the organisation develops.
The Watson Developer Cloud also allows groups to develop cognitive applications catering to their specific needs. These applications can take the form of anything from question and answer services, to tools that evaluate how a message will resonate across the organisation, or provide insight into the characteristics of a user.
Watson is changing the way we support consumers, as well as enterprise. The current model of customer service is already strained by the demands of the millennial generation. This is a generation that has left the days of single shop-floor transactions far behind, and expects a continuing dialogue and high level of support in their purchase experience. Millennials will make up nearly half of the workforce by 2020, and an IBM survey of 1,700 chief marketing officers revealed that 65% of CMOs feel under-prepared to face them.
These are concerns that apply far beyond the technology sector. We are using digital channels for an increasing number of crucial decisions and transactions, from insuring our cars to consulting on our medical complaints. Since the first secure online purchase was made in 1994 (the Sting Album ‘Ten Summoner’s Tales’) online sales have rocketed, with over £100 billion (1.3bn Euros) spent online in 2014 in the UK alone.
Service providers face two problems with a digital model of engagement: the increasing volume of digital interactions, and their increasing sophistication. Despite colossal spending, businesses consistently underperform in handling both of these problems. US companies alone spent $112 billion of call centre labour and software, but a 2013 survey found that every year half of the 270 billion customer service enquiries go unresolved. Consumers have more choice online than ever before, and businesses failing to provide a first-rate customer experience risk losing their loyalty.
Used as a resource in customer service, Watson can help direct enquiries to the appropriate support, by applying its comprehensive store of background knowledge to the interaction.
Watson can improve the way we carry out so many tasks, but applied correctly it is also a truly creative force. One example of this is IBM’s Cognitive Cooking System, which brings together the diverse arts of cookery and analytics to produce culinary innovations such as Indian Turmeric Paella, Baltic Apple Pie and Ecuadorian Strawberry Dessert. All of these dishes contain one secret ingredient, refined by experts in food chemistry and from consumer feedback: data.
The dishes mentioned might be valuable for their novelty at a dinner party; but they also represent solutions to a genuine business need. New ingredient combinations and flavours are potentially extremely lucrative for food service providers and manufacturers, and vast amounts are poured into research and development as a result. Watson’s contribution to this process is twofold: inventive suggestions, carried out at a hugely accelerated rate. A cognitive system can supplement the expertise of a human researcher, as it has no pre-conceptions – we constantly strive to “think outside of the box”, and for Watson, that “box” does not exist.
Despite its commercial prospects, one might be tempted to think of Cognitive Cooking as a niche or frivolous pursuit. But the system also has hugely powerful humanitarian applications – Watson can also explore how locally available ingredients can be combined to produce cheap, appetising, and nutritionally balanced meals.
While Watson’s work in nutritional technology is exciting, this is just one of the avenues that creative computing will explore – differentiation is key in every market. Just as Watson Chef considers culinary tradition, seasonal tastes, and the aesthetic appeal of a dish, a Watson fashion consultant can consider previous fashion icons, the current vogue, and a customer’s existing wardrobe to suggest the next “It” bag or shoe. Where previously we’ve waited for a stroke of genius or a fluke to produce a truly outstanding product, Watson uses data.
Watson plays a crucial role in all of the scenarios discussed thus far, but perhaps none more dramatic than the role it plays in the medical community. Here it is supporting doctors and researchers in expanding treatment options, matching patients with clinical trials, and accelerating novel discoveries into the fight against cancer, starting with leukaemia.
These professionals are currently engaged in a constant game of catch up, fighting to keep pace with the new papers published approximately every 30 seconds. Even considering the huge amount of time they do spend trying to keep up – the average researcher reads an average of 300 papers per year according to the National Institute of Health – crucial data and relationships within these publications are inevitably missed. This approach is also slow and expensive; even the most well-resourced pharmaceutical companies take 10-15 years to bring a treatment to market, and in 2013 the top 1,000 research and development companies spent more than $600 billion annually on research alone.
IBM Watson’s Discovery Advisor is designed to scale and accelerate the process of research and medical treatment, directing Watson’s capability for natural language processing towards publicly available scientific journals. By digesting these with reference to existing research, Watson can identify patterns, or evaluate the viability of a treatment or hypothesis. Brilliant scientific minds are already at work all over the world: Watson brings the work of these disparate minds together, and leverages it as a single expert force.
Fantastic as this prospect is, for too many companies it has remained just that – a fantasy. But partnerships with Watson Discovery are already underway at the cutting edge of business and research: commercially with Johnson and Johnson, and in pioneering oncology research with the three top-ranked hospitals in cancer care (Memorial Sloan Kettering, University of Texas MD Anderson Cancer Center, and the Mayo Clinic). In the medical field Watson is not only improving the quality of care for patients, it is democratising the care experience in the process – Watson provides the same support to the doctors of patients at these pioneering institutions as it does to those of the more than 1 million patients (across 4 continents) of Bumrungrad International Hospital.
Watson is already making a difference, but the problems we must use it to address are equally immediate – in the UK half the population is expected to develop cancer at some point in their lives, and it is this sort of pioneering technology that could help determine their level of care.
If nowhere else, look to IBM’s own strategy to reflect how much of a game-changer this system will be: over $1 billion of internal IBM funding is earmarked for Watson, and a new Watson global business unit has been established. The move to cloud-based hosting and a $100 million investment from IBM has opened up access for developers, and allowed the growth of a Watson-centric ecosystem of over 2,000 organisations and individuals. Although Watson technology is one of IBM’s most valuable assets, Watson Analytics is available as a freemium service. The business sense of this decision is clear for existing users: the power of Watson in practice is its own best advertisement.
When researchers began the project that would become Watson in the labs of IBM, even they could not have realised the magnitude of the work underway. Times have changed: in 2013 Deloitte estimated that cognitive computing is set to be a $50 billion market opportunity by 2018, in the US alone. Leaders everywhere are realising that the technology is here, now, and at work in their competitors and their peers. This is a wake up call – woe betide those who are caught sleeping.
About the Author
David Stokes is the Chief Executive for IBM in the UK and Ireland (UKI) where he is leading the charge in helping organisations across all industries leverage digital technology to transform, both by streamlining their business operations and by creating engaging front office systems.