Everything Oracle

Introduction

One area where Oracle has taken an important step forward beyond the graphical capabilities found in Discoverer is the presence of “Guided Analytics” in the BI Interactive Dashboard and BI Delivers components of OBIEE. Guided Analytics is a major growth area for the future, one that will lead inexorably to “AI Assist”, “AI Cooperative Decision Making”, and finally to “AI Automation” in many areas of business.

Let’s suppose you’re out for a walk in the countryside. If you see something unusual, be it an oddly shaped tree or a curious rock formation, then you’re likely to approach to examine that something in more detail. Studies in the neurobiology of attention show that the key factor required for an event to grab our attention is that it is surprising or out of the ordinary – such an event may well represent either a threat or an opportunity, and in either case it merits a more detailed investigation to determine its significance.

Guided Analytics works on the same principle. A typical example is that of an engineer monitoring a national power grid. When the loading on a certain power station exceeds, say, 90% of its capacity then a red light flashes, and related, relevant information appears on a screen, such as the prediction of future demand, the availability of backup sources, and the time required to bring them online. Such functionality has of course been available for many decades, but it has tended to be hardwired and to be restricted to specific industries where a lack of timely decision making has major financial or safety consequences.

Business Benefits of Guided Analytics

If we consider what the present day diligent manager or executive tends to do in more mundane business areas, then it is to systematically trawl through relevant data sources – principally reports – on a regular basis looking for anomalies in the data that may warrant an intervention. With Guided Analytics that manager or executive can specify in advance what constitute anomalies, and then the system can draw these anomalies to his attention immediately that they arise – the time taken for a certain customer to pay his bills has been steadily increasing, or the annual cost of repairs to a certain item of plant has risen inexplicably. The benefits to business of Guided Analytics are twofold. The first is timeliness in dealing with anomalies – and “a stitch in time” applies just as well in business as it does in domestic affairs when it comes to avoiding unnecessary future expense. The second benefit is that Guided Analytics allows a single individual to manage a greater workload without a reduction in efficiency – very relevant in the current financial climate when businesses are downsizing. If the effort involved in analyzing data “to see if a problem is present” can be reduced, then more effort can be expended on remedying those problems that have already been identified.

Guided Analytics in Dashboard and Delivers

Within a BI Interactive Dashboard it’s possible to customise the display layout and the content based on the information that is being received, and it possible to define a BI Delivers iBot using a conditional request that will send a report to a user when a certain business event has occurred. Exactly how you would use these capabilities within your own organization is something that is worth exploring with your end users, possibly by knocking up a small prototype that provides a proof of concept in a relevant business area.

Guided Analytics can also be used as a means of disseminating “best practice” within your business. The current approach of issuing written guidelines has a pretty dismal record when it comes to improving performance, and, as it also smacks of senior management telling middle management “how to do their job”, it is often met with resistance. Instead, by taking aside a few managers known to be particularly efficient, it may well be possible to extract some information about how this efficiency is achieved: what data sources do they use; what subsets of information do they examine to detect anomalies or to trigger action. For example, a manager might say that when regional sales for a product vary by more than 20% he will drill down and examine the sales for individual stores, to determine if the variation is uniformly spread across the stores, or if it is concentrated within one or two. It’s not difficult to build such behaviour into a dashboard or an iBot so that, in the case of the example, whenever a 20% variation in regional sales with respect to previous values occurs, then a warning appears or a report is generated drawing the fact to the user’s attention. Using Guided Analytics in this manner allows “best practice” to be propagated throughout an organization in a non-intrusive manner.

Additional Channels – Sound and Haptics

But there is more to Guided Analytics than pictures on a screen. Before too long when a customer rings up a telesales operator, guided analytics will add a voiceover into the operator’s headset reminding him of relevant information about the customer’s past purchasing habits, or perhaps haptics build into a mouse or keyboard will provide similar information via the sense of touch. Sound and touch are candidates for additional “dashboard components” of the future, just as vision is today, and indeed in data intensive activities, such as piloting an aircraft, these other senses are already used to assist in prioritizing an individual’s attention on the information that matters most. Now it’s unlikely that operational decision making in most businesses will ever involve the large information volumes and require the rapid responses present when a bird strike shuts down an aircraft engine or the stock market learns that yet another major investment bank has gone bust. But psychologists have found that presenting information through different sensory channels can improve the manner in which it is processed and integrated. In particular, using additional channels can reduce visual overload and may increase efficiency and attentiveness when performing tasks, such as those of the telesales variety, that are largely repetitive in nature.

AI Assist, AI Collaborative Decision Making, and AI Automation

Along with Guided Analytics will come – for better or worse – a sea change in workforce control and monitoring. Monitoring of the productivity of shop-floor workers by software has become routine in many businesses, but management decision making has, until now, been considered too complex a task to be readily quantified and controlled. But with Guided Analytics we are beginning to see the first tentative steps towards involving software in the decision making process.

It all this seems a little like the HAL of 2001 informing Dave of matters that require his attention – “I've just picked up a fault in the AE35 unit. It's going to go 100% failure in 72 hours” – then of course it is. With Guided Analytics we are entering the realm of AI Assist in which software becomes ever smarter and increasingly helps to guide the decision making process. At this stage we are in the foothills of AI where assistance is restricted to presenting relevant information on which a decision can be based. But expect it to move, in due course, to “AI Collaborative Decision Making”, in which the AI system not only presents the relevant information, but also suggests how the user should react to that information – presented as a list of ranked options, together with their respective pros and cons. And, in time, when the software’s preferred option leads to consistently better results than the collaborative decision, then the era of “AI Automation” will dawn when the user’s role will be restricted to that of rubber-stamping the software’s decision, or of occasionally overriding it when unusual circumstances prevail – or perhaps not, “This mission is too important for me to allow you to jeopardize it.”

OBIEE has yet to embrace sound or haptics, much less AI Collaborative Decision Making or AI Automation. But as these technologies are pioneered successfully by small start-up companies that are in due course acquired by Oracle, we can expect all of them to appear in some successor to OBIEE – perhaps, OBBEE (Oracle Big Brother Enterprise Edition). That’s all for now Dave!

Everything Oracle

Home

Everything Oracle