Are you suffering from artificial intelligence (AI) fatigue? Like the words of the song, “every move you make, every step you take, I’ll be watching you,” AI pervades all our activities, from work to leisure, analysing data on every action, from a simple grocery shop to clothing preferences.
While expectations are high, the reality is that true value is not always realised from AI in analytics at this stage of the game. There is no doubt that the potential is huge; however, with typical ‘black box’ approaches, disappointment is also common.
Simply put, a black box is an impenetrable system − most people are more familiar with the aviation term for aircraft data recorders. An AI black box system is one whose inputs and operations aren't visible to the user or another interested party.
While the evolution of what started as business intelligence (BI) has been significant since its inception in the 1990s, the promise of fully accessible, immediately actionable data analytics remains to be fulfilled.
Let’s unpack the steps in this evolution:
First-generation analytics − report-centric: In the early days, a skilled team within IT managed a complex set of technologies that delivered predefined reports and ad hoc responses to business requests for data. Users would formulate questions, submit to a data analyst, and wait (sometimes for weeks) for a response, usually in the form of a new report.
Second-generation analytics – visualisations and dashboards: With the advent of user-driven analytics, business users were given the power to prepare data, load it and interact with it in intuitive, visual ways. Today, many businesses are still functioning in this manner. While the benefits are clear, many lightweight visualisation tools present challenges around governance and scalability and their complexity limits user adoption.
Third-generation analytics – augmented analytics: This is today’s picture where AI is augmenting and enhancing human intuition. This represents a shift toward a fully democratised framework in which users of all skills levels have the right tools to work with data, generate insights and take action immediately.
The promise of fully accessible, immediately actionable data analytics remains to be fulfilled.
The emergence of third-generation business intelligence is being driven by a series of technical developments that have changed the data and analytics landscape. In recent years, we’ve seen an exponential transformation in the volume, variety and velocity of data available, both on-premises and in cloud environments. This requires businesses to have a comprehensive data integration and management strategy.
Of course, one of the most important capabilities unlocking the third-generation of BI is AI. In the context of analytics, AI leverages machine intelligence to provide insights, automation and new ways to interact with data, helping to drive data literacy across organisations.
The next big driver is the fact that today data is spread across on-premises and multiple cloud sites, where companies need to access it, manage it and analyse it.
At the same time, cloud infrastructure has greatly accelerated the ability to scale and is providing the compute power needed to manage and analyse vast quantities of data.
Finally, the arrival of more active forms of BI is enabling businesses to take real-time action based on changing data. Capabilities like alerting, automation, mobility and embedded analytics – all supported by real-time data pipelines – are delivering the power of analytics where and when action is needed.
Now we get to the jewel in the crown and the key to successful AI in analytics today: augmented analytics. It should be noted, however, that augmented analytics has been around for a while, but it now means something much more powerful as the AI doing the augmenting is so much more potent than it was when the term was first coined.
This is an approach that unites the best of machine intelligence and human intuition to speed time-to-insight, surface new and unexpected discoveries, and drive data literacy for users in any role and at any skills level.
While there are niche applications for AI that completely rely on machine algorithms, most complex business problems require human interaction and perspective.
Augmented analytics creates a multiplier effect, where the human-machine collaboration outpaces anything either the human or the machine could do on their own, and that’s not the only benefit.
When people participate in the analytical process, they tend to have confidence in the results – whereas any conclusion that comes fully formed from a black box will naturally raise doubts. Augmented analytics breeds trust, resulting in more buy-in and ultimately greater adoption of analytics and the insights they provide.
In my next article in this series, I will expand on what augmented analytics looks like, its existing capabilities and where it’s going.
Share