One of the many dictionary definitions of the word ‘smart’ includes the phrase “quick-witted intelligence”, which seems as good a way as any of describing what we currently mean by a smart enterprise. Definition is a challenge yet to be met because there is no tech spec for smart. But both the IT industry and its corporate clients share a broad vision of what we are aiming for — an organisation that is fully joined up digitally and capable of realising the benefits of that synergy. That in turn means corporate agility based on the ability to utilise all of its collected data, complemented by relevant external data, in real time. A useful term in the ether now is ‘Augmented Intelligence’, which will help us somewhat limited humans to make better decisions.
Analytics, automation, connectedness
The lines of development towards that vision are almost all predicated on the continuing development of analytics, coupled with automation, robotics (robots and bots), sensors and the Internet of Things (IoT), smart machines and machine learning and ultimately perhaps artificial intelligence. Cloud has supplied the computing, storage and sharing capacities for all of these advances but in many ways it is just the platform of choice. Who knows what alternatives the future will offer.
To be fully joined up digitally is the ambition not just for the enterprise, but as we know for other types of organisations or entities such as cities and towns, transportation systems, governments (local, national and supranational), security or military or whatever. The smart ecosystem certainly includes also individual artefacts from buildings to ships to aircraft to data centres to factories. In fact, the potential and value of smart systems are probably clearer in such examples than in the somewhat abstract ‘organisation’.
Smart cities has been one of the major research streams in IBM’s Research Lab in Ireland and its current director, Dr Wendy Belluomini, points out that the work is more than relevant to the enterprise. “Cities are a great place to start. We have clearly established the value of working with the physical world in real time, harnessing the fairly recent combination of sensor technology, cloud computing, machine learning and analytics.
Real-time move
“Generally in enterprise IT we have moved on from automating a relatively fixed set of processes to the next wave in which real time adaptive learning is the key. We are also seeing large scale sensorisation — across multiple and rapidly growing applications — and the result is dynamic adaptivity in our systems. We are also learning new lessons, for example that the flood of newly available data should in some degree be processed at the edge. A lot of our work in IBM right now is about embedding intelligence at the edge of the network so that the centre is not overwhelmed.”
Dr Belluomini is keen to add that IBM’s original work on smart cities and the physical world is matched by service industry possibilities. “For example, we are working with the financial services industry on the use of blockchain technology. It has serious potential to supersede traditional ‘paperwork’ recordkeeping across a wide range of things and we are working on how to make it faster, more automated and streamlined. In insurance, we are working on how companies can interact with customers in a more automated way. That is not to replace humans with machine learning, but to assist them to provide a better quality service with better information and to make the more difficult decisions.
“But in all service industries and professions there are people spending a high proportion of their working time on tasks they could be helped and supported with by smart systems, especially where there are huge amounts of data. In Watson Health, for example, IBM offers a system that reads something of the order of 10,000 medical journals which no doctor could manage to do — even keeping up with a particular speciality. So the system can assist the doctor to be much more effective by providing all current information to help in clinical decision making. We are talking about systems as assistance.”
Assist not replace
That theme of smart systems of analytics and automation being of assistance to humans is also a through line in the work of Hewlett Packard Enterprise, which uses the term ‘Augmented Intelligence’. “Artificial Intelligence is certainly a hot topic but it is generally misunderstood,” says HPE’s Seán Blanchflower. “People think of robots doing most of what humans can do but we see it as assisting people to make the best decisions. We have been working in this field for about 20 years and we see it all as beginning with the analysis of Big Data. In my group currently, we are focussing on what we call Human Information, which is designed to be consumed by human beings.”
The early days of computing, he points out, involved the entry of data in narrowly defined databases that, in a sense, only the machines could make sense of. In the 80s with the advent of email and especially in the 90s with the arrival of the Web, people began to produce computer content, from documents to video. “Now we are talking about huge volumes of unstructured data that only humans can really understand — think of camera streams or audio recordings. That’s what we specialise in, because today about 90% of the information in our systems today is unstructured human information, not tidy databases.”
In that context, Blanchflower and HPE see the impact of analytics and machine intelligence as optimising the workforce rather than replacing it. He gives the example of a driver in today’s world of logistics and distribution. “In the past a driver would have relied on past experience, maps in new places, intuition, observation to make decisions about routes, traffic flow and so on. Today, a combination of telematics and route optimisation data add to the driver’s judgement in a kind of hybrid process that maximises efficiency.”
Ltd
