The term ‘Augmented Analytics’ was only coined in 2017 by Gartner. But in many ways what it describes is a logical and perhaps inevitable technological progression – the marriage of Big Data analytics techniques with Artificial Intelligence (AI).
It wasn’t that long ago that everyone was talking about the revolutionary impact Big Data was having on business and industry, arguably even establishing data science as a key value-generating discipline.
Before Big Data technologies emerged, analytics played a fairly descriptive role in business reporting and business intelligence (BI). It could tell you, for example, how sales volumes or profits compared to comparative periods previously, and reveal patterns that could be useful in planning and decision making.
But what conventional analytics was not geared up for was delving into the whys and wherefores that explain the patterns. As organisations have gone through an ongoing process of digital transformation over the past decades, data containing these kinds of deeper insights has become more and more available. But conventional algorithms struggled to handle the vastly increased volumes of data being generated, and they were not designed to compare and contrast findings from data that came in different forms from numerous different sources.
This is why Big Data analytics was a game changer. Now, organisations had the analytical power to delve deep into the very largest data sets, identifying patterns from all their different digital systems to give detailed intelligence not just about what was happening, but why. This has proven invaluable to many businesses throughout the world.
Yet getting to this point of detailed, granular intelligence is not straightforward. Big Data algorithms are great at working with incredibly large datasets, but they still need to be prepared. If you are using data from multiple sources from across the organisation, first of all it needs to be pulled together, collated, ‘scrubbed’ and put into a consistent format that the algorithms can use. Then, when analysis is completed, the results are still only given as a list of numbers. To be turned into meaningful intelligence, it needs interpreting and turning into the types of charts, graphs, narratives and the rest we are familiar with from reports.
This is why data science roles have become such sought-after positions across the world of business and commerce. Big Data has opened the door organisations being able to gain huge value from their data assets in the form of detailed, actionable BI. But working with data in a way which allows this to happen remains a highly specialised skill.
Augmented analytics – that is, Big Data technologies with an added boost from AI – provides more of an end-to-end data handling solution. To put it simply, AI can now be used to perform many of the tasks that data scientists have so far had to take on. Machine Learning (ML), for example, enables a level of ‘smart automation’ capable of collecting, collating and cleaning diverse data sets ready for analysis, improving its ability to compare different types of data over time. Natural language generation (NLG) and visualisation technologies translate information contained in data into a form humans are used to dealing with, whether it is textual explanations and descriptions or charts and graphs.
Moreover, Augmented Analytics is able to drastically reduce the timescales involved in turning raw, unfiltered data into actionable BI, from hours or even days into something close to real time. As with most types of software automation, it improves accuracy and reduces the likelihood of human error, particularly in relation to data preparation.
When you read around the benefits of Augmented Analytics, you frequently come across phrases like “it democratises data analysis”, or introduces “self-service analytics”. In other words, the fact that AI-powered automated analytics can perform many of the tasks data scientists were previously required to perform is widely seen as a benefit, especially in a climate where skilled data scientists are at a premium.
So, could we be at a point where the demand for data scientists sparked by Big Data is about to wane with Ai-driven automation taking over?
There is a clue in the fact that Gartner chose to label this new generation of technology ‘Augmented’ Analytics, not automated. We should see AI-powered analytics as supplementing the work of data scientists, not taking it over, and perhaps ushering in a new approach to data science altogether. In particular, if AI takes on a lot of the complex, time-consuming technical tasks which data scientists have previously had to dedicate most of their time to, this creates new opportunities for data professionals to look beyond the algorithms and pay more attention to what they are being used for.
Perhaps the biggest opportunity of Augmented Analytics is that it opens the door to addressing a long-standing bugbear of many IT and business executives alike – a lack of mutual understanding between the people who know their data and the people who need to know what the data tells them. With AI onboard, we will maybe see data science shift to take on more of a strategic role in identifying where data-led intelligence can have the biggest impact.