UDEK - An article written 12 years back - before BIG DATA became BIG!

ENTERPRISES NEED UBIQUITOUS DECISION ENABLING KNOWLEDGE (UDEK)

Information has become a self-feeding loop for the Enterprises. The realization that - The more you know the more you need to know - has yielded unprecedented hunger for more data and information at all levels in the Enterprise. However, this information and data explosion yields a directionless collection of data at various levels in the Enterprise. This leads to Information de-coherence and leads to high volumes of collected data lying unused. The data could have been put to a proper usage in creating Decisions at various levels in the Enterprise. There is a requirement to convert the incoherent data into actionable knowledge we call Decision Enabling Knowledge (DEK).

To create this DEK, Enterprises need to embrace and combine variety of quantitative and soft-computing techniques from multi-dimensional areas of Operations Research (OR), Cybernetics, Artificial Intelligence (AI), Statistics, Applied Mathematics and System Analysis. This article describes various techniques and their applications for converting mountains of data into genuine insights for transforming Enterprise business.

  

Data Explosion to Information De-Coherence

A large living Enterprise generates huge amount of data in a single day. This data may be about sales, supplies, inventory, customers, employees, finance, strategy, operations, competitors, markets, and wholesome mixture of miscellaneous activities. Presenting a single holistic view of data to create optimized decisions regarding various activities and at various levels is a major challenge faced by large Enterprises.

Data need to become a Common Operating Picture (COP) at all levels of the Enterprise. In reality, however, COP is usually not achieved leading to pockets of data and information residing in various places within the enterprise. These pockets are usually disjoint and give piecemeal information. Further the information needs to be processed for generating knowledge required for making decisions. As described in the May 2002 issue of EAI journal (refer to article titled Creating Next Generation Enterprises) the Decision Enabling Knowledge (DEK), as we call it, is usually generated first just before the decision is to be made. This involves reaching to relevant information in multiple places, collating and processing the info nuggets and converting these into relevant knowledge for the decision maker. This process eats up a chunk of time that may be critical in real-time decision-making that the enterprise needs. The delay in processing may be the difference between a win and loss for the employee and in turn the enterprise. Also, due to lack of time and non-availability of DEK, general tendency is to make decisions based on intuition, judgment and experience. This may lead to decisions based on dated information thereby to disastrous results for the enterprise sometimes. Despite the needed knowledge being available somewhere in the enterprise, it fails to reach the employee at the time, at the place, and in the way the employee needs.

Decision Making – Requirements

When we look at how a decision is made – we can see five steps. First step is the assessment of the situation including where we stand with respect to the environment. An enterprise is a living entity. It exists in a remarkable dynamic environment consisting of the consumers of enterprise products, competitors producing the same or similar products, suppliers of unprocessed raw material that go into making of product, employees, internal enterprise dynamics, etc. Assessing a situation keeping the interest of the enterprise as the objective becomes a Herculean task sometimes. This assessment involves collection, collation, filtering and classification of data. Next step is the explanation of why are things as they are. Events and trends have an interrelationship that result in great variations in outcome of their interactions. Accurate explanation of why are things as they are becomes an important step in any decision making. Third step in the decision making is the prediction, projection and forecasting. This involves analyzing what will happen if the trends continue as they are. The fourth step is generating options regarding what all can be done to align the situation as per the requirements of the Enterprise. Last step is choosing the relevant option or combination of options to be executed.

Every decision that is made in the enterprise undergoes these five stages. However, the time given to each of these decision steps varies considerably based on the value of the consequences of an incorrect decision in the mind of the decision maker, time available for decision making, accuracy of the data available at each steps. Figure 1 illustrates the flow of these five steps.

The requirements at each of these decision making steps are shown in Figure 2. As is evident the requirements of decision making starting from data collection to optimized choice of option are enormous. It must be emphasized that meeting these requirements for every decision in the enterprise should be the endeavor of the technology adoption strategy in the enterprise. 

Can Technology Help – Creating Decision Enabling Knowledge

Science has provided many formal and quantitative techniques to meet the requirements of each of these steps in Decision making. . These techniques have been carved out of years of research in the fields of Artificial Intelligence (AI), Operations Research (OR), Cybernetics, Linguistics, Neural Networks, High Level Simulation, Modeling and System Analysis. If technology can amalgamate these techniques to provide processed information along with heuristics and rules to assist each step of decision making at all levels of the enterprise we can say the huge investments being made in the technology will be justified. This will lead to Decision Enabling Knowledge (DEK) becoming available to employees at all levels of the enterprise. Figure 3 gives some of the techniques that can be used for generating DEK at each step of the decision making process.

Although many commercially available software packages exist based on these techniques, there is a need to integrate these techniques into a comprehensive whole, so that enterprises can leverage this body of knowledge for generating value. An integrated architecture that combines all these techniques into an end-to-end decision making engine is the need of the hour for large enterprises. Although it is possible to have such an engine, no commercially available package offers this. It is left to the enterprises to decide on the extent to which they want to adopt this. Given the state of affairs many large enterprises are not even aware that these techniques can help them in a major way.

Ubiquity – The Umbilical Cord to the Employee

Creating DEK is not sufficient. It should reach the decision-maker at the place and time, the decision-maker want. Here continuous connectivity to the enterprise infrastructure is required. This will involve providing the decision maker with the facilities to reach the DEK lying in enterprise infrastructure. With the advent of new and powerful mobile devices it has become possible for the remote employees, tele-workers and employees on move to be connected to enterprise infrastructure. These devices are, however, still limited in capability compared to desktop machines. With the advances in miniaturization, longer life batteries, sharper displays, higher on-device storage capabilities, the mobile devices will become as powerful as the desktop of yesterday.

The key question however, remains – what should be available to employees on these mobile devices. Here we propose the DEK combined with personalization and location dependent systems should be made available to all employees on these devices. The employees should be able to access the DEK at the place, on the device, at the time and in the situation they are in. While designing such Ubiquitous Decision Enabling Knowledge (UDEK) systems, Enterprises will face stiff challenges. The technologies, the infrastructure and employees inertia may be hindrances. However, it is with UDEK as the goal Enterprises should invest in technology.

Technologies for UDEK – EAI, Web Services and BPMS

Creating UDEK at all levels in the enterprise requires integration of multiple systems and units in a comprehensive, dynamic and flexible infrastructure that is live at all times and is available to all. Producing such pervasive infrastructure will require adoption of multiple technologies that are becoming available now. With the advent of Enterprise Applications Integration (EAI), Web Services and Business Process Management Systems (BPMS) at the enterprise level, producing UDEK has become possible. With all said and done about these technologies, their task is to assist in making optimized decisions by providing collated, cleansed and correct knowledge relevant to the decision issue at hand.

If enterprises are talking of EAI, WS and BPMS as separate technological options, there is a surety that enterprises will count their trees but miss the forest. The strategic goal of creating UDEK is what will generate a cohesive flow for achieving maximum ROI and minimum TCO. If DEK has to created, applications need to talk to each other may be trough web services, multiple applications need to be integrated to perform relevant functions, and all these applications need to be woven into dynamic, flexible and agile business processes. The infrastructure thus created will need to generate and propagate DEK to all levels at the enterprise.

Conclusion

This article enumerates various techniques for creation of Decision Enabling Knowledge (DEK) at all levels of the Enterprise.  It combines multiple techniques from various fields such as Operations Research (OR), Artificial Intelligence (AI), System Analysis, etc, into a loosely coupled framework to provide genuine insights based on analysis of disparate data for solving various real life Industry problems. It is proposed that Ubiquitous Decision Enabling Knowledge (UDEK) should be the goal with which the enterprises should adopt the new technologies that are coming up from multiple directions. Unless enterprises work towards UDEK the technology adoption will keep on creating problems that in the long run will create more problems rather than solving the existing problems.