Always rapid, the velocity of technology change is speeding up. This is particularly evident in the artificial intelligence (AI) space. While it took around 16 years for AI to reach human competence in handwriting recognition, in only two years it exceeded 80% of human capabilities when it comes to generating code.
For companies, the key issue is their ability to keep pace and invest in future-proof systems and tools that can adapt to constant evolution.
The field of business intelligence (BI) offers a unique view into how tools are thriving or just surviving due to the rapid evolution of AI.
The challenge when implementing BI platforms
As we all know, BI has been around for decades, and many companies have taken advantage of the added capabilities offered by BI.
But there is a large installed base which typically is a mix of disjointed tools aimed at three distinct areas: data preparation, data science and finally business analytics.
This heterogenous environment creates complexity, which in turn increases costs because expensive specialist skills are needed, particularly in the data science/AI arena.
And, of course, the more complex anything is, the slower it moves. The consequence is delayed business decisions, resulting in negative business impact.
A key disadvantage of traditional BI is that it is something of a black box: the decision-maker just gets an insight and/or a recommendation, but has no way of seeing and assessing the underlying data on which it is based.
This combination of BI, AI and machine learning forms the basis of “augmented analytics”, and it’s the future for business.
Unsurprisingly, managers are uncomfortable using BI to make decisions because they just can’t assess for themselves whether the underlying data is convincing or not.
A further disadvantage − and this might be the most significant of them all − is that to get the full benefit of BI, users have to be able to understand the data parameters and write the code needed to generate the insights they want. Those who lack these skills have to wait for experts to write the queries they need.
In short, however useful some of these insights or recommendations are, they often come too slowly, and are not available to everybody.
The power of democracy
The typical BI environment is simply not fit for purpose in today’s milieu, which prioritises rapid decision-making based on sound inferences from multiple streams of data, both internal and external.
A key point is that the insights need to be available at the point of need, not isolated in some expert silo, and they must be provided in near-real-time. In turn, that requires making it possible for managers who are not data experts to develop their own queries and analyses, without the need to do any coding.
In similar vein, the insights generated need to be presented in such a way that the non-expert user can quickly understand the implications − they need to be communicated as a story, in plain English, not as a set of graphs. Follow-up questions must be easy to ask, and similarly the answers must be understandable.
In practical terms, it’s important to achieve this while leaving the existing heterogeneous environment of BI tools in place. There’s no time or budget to put in a “new and improved” system”.
The answer is a universal platform that can incorporate all these tools. AI and machine learning can be used to do this, and AI will also play a role in providing explanations for the user.
In a visualisation, for example, the user would now be able to pick any data points and the program will drill down into the data model to provide rounded explanations, explain anomalies, generate descriptive and diagnostic analysis, display the relevant data, and clearly describe the reasons for underlying differences between any two data points.
This combination of BI, AI and machine learning forms the basis of “augmented analytics”, and it’s the future for business. It enables the decision-maker to quickly form a view of the underlying logic of the insight or recommendation. He or she can then decide whether it is valid or needs further investigation based on his or her own experience and knowledge.
It’s a good example of how this kind of technology can be used to augment human capabilities and magnify the quintessentially human ability to make counterintuitive connections between facts, making decisions that can potentially enhance competitive-edge and improve the business.
Overcoming fear of AI
We’ve all seen the flood of articles in the media about the perils of AI, with even tech titans like Elon Musk arguing it needs to be regulated. Be that as it may, there’s no doubt there is considerable unease about the impact that AI will have on people’s jobs.
It’s a big topic, on which one could have various viewpoints, but the smart money seems to be saying that, used wisely, AI can enable huge gains in productivity because it amplifies human capabilities.
For that to be possible, it needs to be available to, and useable by, everybody in the company that needs it, when they need it.
Platforms that can pull all the existing BI tools together to deliver understandable insights and recommendations, and that provide explanations for how conclusions were reached, will be able to do just this, empowering employees at all levels to make better decisions rapidly and with a high level of confidence.
Making AI easy to use, and demystifying what it does, will help in reducing employees’ fear of the technology, and make them more likely to use it. The result? A sharper, more competitive company.
To get a sense of what can be achieved, consider South African fashion retailer TFG. It used an augmented analytics platform to consolidate the various pockets of analytics that previously existed across the group, each associated with different systems.
TFG processes around 20 million records daily just for its stock-in-hand inventory, says a senior department manager − too much for its existing BI tools to handle. Using its new analytics platform, TFG is now able to achieve a single view of data pulled from multiple business streams, all using a variety of products from different vendors.
Financial reporting is also much more accurate and timely reports that took 3.5 days to complete now take under 15 minutes. These reports give buyers and planners much deeper insight into outcomes and are also used by executives to inform business strategy.
There are numerous other benefits, as can be seen in the published case study.
Data is the new oil; we are often told. True, but mining it productively now relies on AI and machine learning to realise its full potential to lift the performance of the entire business significantly.