Whether business needs machine learning, if business analysts can replace the decision-making process with a person, with what stereotypes one encounters in discussing artificial intelligence – the answers to these and other questions are provided by Wael Elrifai, Sr. Director of Pentaho Enterprise Solutions EMEA & APAC, Hitachi Vantara.
– In your view, in which areas the use of machine learning and predictive analytics can bring the maximum effect?
– Predictive Analytics refers specifically to Supervised Learning (learning to predict outcomes) while Machine Learning can encompass the field of unsupervised learning – trying to find structure in data (e.g. clustering). I can’t think of a field that won’t benefit from these technologies; whenever we’d like to predict the future, simulate potential alternative actions, and optimise outcomes, machine learning will bring great benefit to companies, and to society.
– How do you see the future of business intelligence technologies? In which direction will they be developed?
– I think this is a question of semantics, business intelligence is generally a “look-back” technology, understanding how we’ve performed and operated in the past. Increased automation is likely to reduce the amount of hand-coding required to gain the look-back insights.
– What is the role of Business Analytics in the processes of digital transformation of businesses?
– Data and analytics should be at the core of every digital business platform. It is the heart of bringing together people, things and IT Systems. Companies must think big, start small and scale fast in order to compete. The key to enterprise transformation lies within data. Big data, IoT and machine learning are already changing the world of data from one that is siloed to one where data is connected, open and fluid. Established barriers across infrastructure, operational technologies (OT) and IT, as well as across the broader world of societies and populations, are being blurred. Ultimately, no company will be able to navigate through this uncertainty without having the capabilities to build data-driven businesses using business analytics.
– What business areas/industries are/will be more advanced and actively using data analytics technology, why?
– Interestingly enough, it’s the more traditional manufacturers and service industries, especially financial services, which are the most advanced users and early adopters of data analytics technologies. One reason is that all these players are being threatened by new, disruptive market forces and competitors. There is a big race going on in the automotive sector, for example, to produce smart, safe, energy-efficient cars of the future that are connected to insurance providers, roadside assistance and other services. All the systems in these cars are defined and connected by data.
In another example, traditional banks face grave threats from a new wave of lean, agile fintech companies that provide exceptional online experiences and competitive rates. At the same time they are saddled with expensive, decentralised legacy infrastructure and myriad database, analytics and other applications including BYO devices and software. This legacy data is too valuable to discard, but highly expensive and complicated to manage. These and other reasons explain why so many financial services are investing in data analytics technology like ours.
– Adepts in the data science approach argue that compiling predictive models can ensure bigger profits for companies and bring more new customers than usage of different marketing tools. In reality, how data analysts are able to transform a business model?
– The most obvious change will be a movement towards outcome-based-pricing or XaaS (Anything-as-a-Service) as companies improve their ability to predict outcomes. Data analysts can analyse the technical feasibility of making the required predictions but there will still need to be robust business change programmes in order to really benefit from these new predictive models.
– Of course, a business manager must take into account the results of the analysis, but he/she can make decisions guided by his/her own intuitions. Is it possible to say that the development of big data is moving towards automation of decision making? Can business analysts replace the intuition-based decision making of a manager?
– A data analyst can make predictions, but only someone well-versed in the business will know what business transformations will be required to actually gain value from predictions. I do believe that there will be a great deal of automation; all 3 previous industrial revolutions came with dramatic automation, and a movement of human labour up the value chain.
– Are people the only users of big data analytics results? Given the development directions of IoT, would it be possible that computer systems will produce and then use themselves the results of such analytics?
– I’ve seen many such cases where systems are both the producer and consumer of data. The simplest are algorithmic trading systems but Reinforcement Learning and Generative Adversarial Network based systems regularly use automated feedback to improve themselves. In fact, all supervised learning methods would also benefit from being consumers of such data – think of Google’s PageRank approach – the more someone clicks on something as the chosen result to a search, the more likely it is to move up the search results list.
– Let’s talk about artificial intelligence. Do you face some stereotypes and misunderstanding of this technology from the customer side?
– The biggest misunderstanding is that it’s automated… that we just feed data in and it comes out with predictions. I think this comes from the unfortunate name, “Artificial Intelligence” where we expect some kind of human-type cognition in these systems.
– How will AI technologies influence business?
– AI is a huge field, Machine Learning is part of that field but it covers Game Theory, Collective Behaviour, Evolution, Adaptation, Network and Systems Theory, Pattern Formation, and more. A tremendous amount of the current research has direct business application in solving prediction problems, or making entirely new discoveries. Right now we’re pretty strong on the former, and only just dipping our toes in the ocean of the latter.
– What is your opinion on the growing role of artificial intelligence for business and on the ethical side of this innovation concerning the human factor?
– Machines are not likely to replace humans in the foreseeable future, or ever. This is because tasks that humans are really good at are the same ones that machines are terrible at – and vice versa. For example, unlike machines, humans are bad at making accurate, repetitive calculations. Computers, however, are notoriously inaccurate – even great image recognition algorithms fail more often than not – when it comes to say, recognising the presence of a cat in a picture. A healthy four-year old child, on the other hand, will always accurately recognise a cat, regardless of the cat’s age, size or breed, or whether the photo features the cat hiding on the couch, under a table, in tree, etc.
Having said that jobs will definitely change and at an individual level, there will be some short-term employment shocks. However, history tells us that as productivity increases, people’s health and leisure time is also increased. The social contract we’ve implicitly accepted as a society is that we encourage companies to develop disruptive technologies, and we provide government assistance to those that face a dramatic negative impact; either through job re-training, benefits, social housing, and other transfer payments.