Former CIO
Tata Power
Were you able to watch the recently held Paris Olympics? Witnessing the 100 mts sprint I was, as always, overawed by the physical fitness of the runners. Every part of the body, small or big, irrespective of its location in the body or its apparent importance or strength was in complete synch and aimed at achieving the one and only goal of running the race. Quite naturally, when every constituent component of a complex entity works in harmony, its full potential emerges. This synchronous working, I realized, is dynamic, with each part functioning to satisfy the requirements of the whole, at that point in time. It is this dynamic relationship that is perhaps causal to the very high level of performance of these athletes during the race.
A similar situation, I reckon, obtains in the functioning of any organization. If all components of an organization operate in dynamic synchronicity with the other parts towards a common goal, the entity will surely deliver at its peak. For officials charged with the responsibility of managing any organization this assertion holds substantial promise but the real question is how does one achieve this united functioning of a complex entity like an organization?
Need for Integration
To ensure that an organization, with its myriad component parts, operate in harmony as a unified entity requires a high level of functional and operational integration within. In a human body the complex network of body parts and components ensure this integration but in an organization this integration must be engineered. Quite naturally this intimate integration cannot be achieved through any one means but requires a multi-pronged approach. It requires a certain level of commitment from the leadership, apposite goals, a clear plan for the execution, involvement of employees and so on. Among this motley host of factors (for integration) a significant enabler is the function of Information Technology and its intelligent usage and that is the focus in this article.
Integration As It Was
In the pre-computer days documented work systems and procedures used to facilitate the process of integration as did company manuals etc. These processes would be designed based upon the learning gained from experience and accordingly work systems and procedures would, be designed and documented. Adherence to these articulated processes would ensure a level of coordination and integration of actions and operations across functions and locations. However, the integration was much limited in scope, nowhere near the current day requirements.
Automation of the Integration Process
Following adoption of computers in organizations things started to change. Those of us who have been around long enough would recall how the process of automation commenced with bespoke applications like share registry, payroll, inventory control, basic accounting, etc. As these slivers of organizations were being automated, expectations were growing for the computerization of larger portions of the organization. Soon the realization dawned that partial computerization was just not enough. The initial steps in the larger integration efforts came in the form of Manufacturing Resources Planning followed by Material Requirement Planning and finally Enterprise Resources Planning (ERP). The ERP was the answer to all our prayers of integration as this helped automate several critical functions of the organization.
Subsequently, newer technologies such as IoT, Edge Computing, further augmented by other technological advancements like Social, Mobile, Analytics and Cloud (SMAC) followed by Ecommerce and more recently Digital Transformation, have enhanced the ability to integrate other parts of the organization. However, while systems moved towards overall integration the logic of integration, has not altered much; it has been largely rule-based and comparatively static, and this has prevailed over time. Let me elucidate this point further.
Nature of Automation
As mentioned, the attempt at integrating operations and functioning of an organization, started off with design of systems and work procedures. Often these work processes were documented in procedure manuals and members located in different parts/functions/departments of the organization carried out their work as per these manuals. Let us consider an example of material receipt. The Purchase Department places an Order for a particular material to a vendor. On receipt of the goods these are unloaded in a receiving yard/store and the Inspection/ QC function need to check whether the supplied material is as per the order placed, is the quality and quantity and price same as what was agreed upon and so on. Having satisfied themselves of the correctness of supplies they will now shift the material to the stores and update records in the stock registers. The material is now available for consumption within the organization.
This simple instance highlights the way in which an organization worked earlier, in a coordinated manner for a common goal. Since these were essentially manual processes dynamic rectification of errors or any change in the process required because of external factors was possible and could be easily incorporated. As mentioned earlier, in the initial days of automation, bespoke applications were the norm and parts of these work processes were on computers while other parts were on paper. With time and maturity these slivers of the organization where automation had been initiated, grew and larger portions of the organization were addressed by computers. The culmination of all of these was the ERP.
The Logic of Integration
Computerized methods utilized to integrate functioning of the organization were based on the existing work processes or was designed afresh keeping in view requirements as had been identified, from experience. This sequence of steps was fed into the computer system along with all possible known variations or requirements of the process. Essentially, it was the manual process that was automated, using technology available at that point in time, and comprised rule-based applications.
All thinking and intelligence were outside the automation process and were factored in during construction of the integrating mechanism. The logic was simple “if so, then…..” and as is obvious this was a limiting factor since the integration could not address any scenario that was not thought of or experienced earlier. Since our design was based on whatever had happened earlier, we had effectively, prepared the system for the past. Ordinarily this worked quite well primarily because by and large the past would replicate and expectations were appropriately pegged.
However, today the stable world for which this approach was apposite, does no longer exist. Over time the nature and density of interconnections between factors separated both temporally and spatially has increased by leaps and bounds making any operating organization vulnerable to frequent change. The current method of integration assumes fixed relation between factors and cannot be altered on the fly as may be justified, which is in complete contrast to the requirements of the dynamic nature of today’s reality. Such limitations of the current integrating mechanism result in certain critical requirements remaining unaddressed.
To summarize, today’s techniques, within an organization, to integrate multiple systems typically require the following:
a) Pre-defined rules on which to integrate, and
b) Pre-defined translation logic to translate the request and response into structures which the systems individually understand.
If newer systems, within the organization, need to be integrated or newer integrations within existing systems are required then the detection of such requirements typically arises post facto with manual analysis and the latter playing a very important role in facilitation of such integration.
Inadequacies in the Current Functioning
Consider for instance, an unexpected development in the environment of an organization that could mean disruptions in certain processes. Ideally could the internal systems of the organization quickly adapt the relevant operational or monitoring systems to the altered circumstances with no manual intervention? If these interconnections were not factored in earlier the system, in the current integration methods, it will not be able to draw up consequent modified action plans and the objective of integrating all components of the organization will not be achieved. If one considers how the human body works it is evident that the dynamic integration alluded to earlier, is existent and operative and that constitutes the true strength of a human being. Such a positive outcome will be available for organizations that can integrate similarly and herein comes the possible role of Artificial Intelligence (AI).
Nature of AI
AI holds the promise to ameliorate the difficulty discussed above. Since AI first came in the form of Chat GPT it impressed all of us and a surfeit of suggestions as to how it could be utilized by corporates began to surface. In this case, in the context of dynamic integration, the basic idea behind introducing AI is that it goes well beyond standard programming and brings to play newer and more sophisticated capabilities. The effectiveness of AI is improving by leaps and bounds almost on a continuous basis and holds enormous promise for much higher levels of performance.
In the early days manual achievement, of certain tasks, was superior to that of AI but the gap has been diminishing with each passing day and today AI’s performance is better in many of such tasks. These enhanced capabilities can be utilized for achieving the dynamic integration that is currently absent and thereby ensure superior performance. Such application for integration, in this case, must be a suitable combination of the basic capabilities of the AI application augmented by additional programs or super imposed small language model addition that will specialize the system for the specific functional/local requirement.
A Possible Application Area
A possible area of intervention using this AI integrated systems is that of the IT Security function. In this absolutely critical area of functioning, the current trend is to put together a set of applications that individually protects different aspects/parts/sections of the overall system. The security of the organization is best served, not by a collection of individual bespoke applications but by an integrated system based on a holistic appreciation of the work, the culture, the context, and other similar characteristics of the organization in addition to technology-based solutions. There is therefore scope for an underlying unifying system to enhance efficacy and reduce the risk associated with the security system.
Acknowledgement of the unified nature of protective and preventive security systems brings the realization that this can only be served by an overarching cyber security architecture, systems, and data, that intelligently integrate the currently diverse bespoke applications. A few instances where such integration may be invaluable are in identifying unusual patterns of behavior in a large system or in simulating a security attack on a large network of computers and users or, for that matter, in effectively handling zero-day attacks all of which requires the collaboration and integration, of several security and other applications.
Conclusion
The best performance of an organization emerges when the entire entity works as a unified body. Such unified functioning requires an intimate level of integration between systems, much like the human body. It needs to be dynamic in nature and function under a variety of circumstances. Our current integration methods, that have been passed on to us over time, lacks these two abilities of dynamic integration and functioning under unprogrammed conditions.
Hence integration that can be achieved through these rule-based programmes is severely limited. AI with its learning abilities holds enormous promise in building integration capabilities. The present approach to utilising AI has been to use it for specific function but rarely has it been viewed as a candidate for achieving high level integration and this article highlights this niche application area for AI.
