Automation and Orchestration for the 21st Century IT Organization

By December 15, 2017AI and Machine Learning

20th Century IT Practices Are No Longer Sufficient

The information technology that fuels the digital economy is energizing the modern enterprise and redefining how companies manage their businesses. Forced to move at hyper speed, many businesses are scrambling to keep pace with constant relentless change and struggling to confront an unfamiliar technology landscape and new complex decisions for which there are no precedents.

Everything is Hyper Connected

Hyper-connectivity between people, processes, data, and things (as in “the Internet of Things”) is challenging traditional data management practices and processes. Once siloed by lines of business, data repositories are being connected and aggregated, and integrated business platforms share data across different enterprise business functions and multiple user groups.

Rich Decision-Making Context

Forward-looking organizations are deploying data analytic tools and experimenting with machine learning technologies in a quest to turn vast amounts of data collected from across the enterprise and the extended supply chain into insights that drive better business outcomes. But aggregating and synthesizing information from multitude of data repositories that employ inconsistent semantic models is no small feat. is challenging existing IT systems and tools. Furthermore, the growth in connected communities of information producers and consumers raise ongoing concerns about data security, privacy and ownership.

Business Agility is the Key to Survival

Survival in the digital economy requires great agility to pivot and switch tracks in response to an environment that is constantly changing. It requires ability to deal with unprecedented levels of organizational complexity by orchestrating diverse functions and resources that operate at speeds of business, and can adapt quickly to new events while hiding intrinsic complexity from users and managers.

Traditional IT automation efforts focused on enabling lean operations by taking over repetitive routine and tedious tasks such as testing, provisioning and configuring, routing, and file maintenance. As a result, IT saw automation facilitating faster response to problems, reducing human errors, and improving allocation of human and IT resources.

But for the more complex IT management tasks, automation alone isn’t sufficient.

Building higher level of agility into the organization means advanced mechanisms to sense and adapt to changes accurately, quickly and confidently. These are mechanisms to integrate, orchestrate and synchronize a diverse and complex set of autonomous, semi- autonomous, and manual processes, and a diverse portfolio of self-managing computing and human resources that enable them through complex enterprise workflows.

Forward-thinking companies understand the need to create higher capacity of data-driven agility. They are restructuring long-standing organizational structures and traditional workflows to put high quality decision-making data in the hands of line-of-business users. Rather than restricting access to select few IT super-users and data scientists, savvy organizations are implementing agile self-service models that make information is available in a useful fashion, anywhere, anytime, and provide rich and timely multidisciplinary decision-making context.

The Promise of Machine Learning and Artificial Intelligence

Artificial intelligence technology offers the next step in building automation and orchestration systems, and evolving them into autonomic systems. In contrast to traditional methods of developing automated systems, in which existing knowledge and methods are formalized algorithmically, AI technology can learn from experience, couple this empirical knowledge with institutional policies and best practices, regulations, and enterprise data, and continually adapt and respond accurately to new conditions as soon as they are detected.

Artificial intelligence technologies give organizations the ability to digitize, formalize, and automate a wider variety of tasks currently performed by skilled human experts. For instance, machine learning-based cognitive assistants that learn from the organization’s extensive help desk experience, monitoring agents that provide early warnings before applications or infrastructure components fail, fraud detection system, demand forecasting, and similar knowledge-intensive applications. This class of AI applications can result in reduction in operation cycles, higher predictive accuracy, and increased customer satisfaction.

The Problem with Data

Machine-learning enthusiasts paint a grand vision of fully automated AI-based systems that learn continually by analyzing vast data lakes and deliver superior decisions in response to always-changing business needs. While the concept is very promising, organizations should be aware of the limitations and potential side effects of unsupervised automated self-learning.

Institutional data repositories may not incorporate sufficient range and diversity of patterns, and often include invalid examples such as bad habits and poor decision patterns that have lingered in the organization, undetected, for years. Worse, enterprise data, and, more so, external data such as social media, is very likely to include biases, whether intended or not. It’s easy to understand how datasets of human decisions may include cultural, educational, gender, race, and other biases, which will be propagated by self-learning systems and can result in incorrect actions and harmful discrimination.

Granted, not all automation requires complex artificial intelligence technology. In fact, certain automation should employ procedural deterministic methods that are easier to build, test and maintain.  But AI can be highly effective in detecting and responding to events and patterns never seen before, and offer a superior ability to augment human performance via highly dynamic decision-support information that improves over time.

AI-Driven Autonomics and Orchestration Platform

AI projects can be deceiving. Small experimental systems and early prototypes can be very attractive and appear just short of magic. But deploying an enterprise system can be challenging, having to deal not only with scale, but also with biased and skewed training data, and user culture and resistance to change. Deploying an AI system that is safe and reliable is a tall order for any organization. And finding the necessary expertise and human resources isn’t that simple either.

Although excitement over the expanding role of artificial intelligence and related advanced technologies in many products and businesses is luring many professionals to develop relevant skills and pursue careers in big data analytics, AI and machine learning, qualified individuals and consulting groups with necessary expertise and proven implementation experience are in short supply.  IT organizations will struggle to staff teams with the talent required to build and deploy enterprise-strength cognitive systems within promised budgets and timelines.

These observations are not meant to discourage organizations from embracing a strategy to adopting AI in their businesses. Quite the opposite—cognitive technologies can be highly effective in reaching higher levels of IT autonomics and orchestration, as well as to augment human capabilities in areas that cannot, or should not, be fully automated.

However, organizations must understand the technology and the application development lifecycle, and be familiar with the associated risks and side effects. One way to reduce risks and have better control over the development effort is to use a platform that encapsulates technology and process building blocks and integrates them into business process workflows.

Obviously, a platform-centric approach does not obviate the need to develop in-house expertise. The purpose of the platform is to make these scarce resources more effective by focusing on implementing a business value-based roadmap rather than be squandered on creating basic technology building blocks from scratch.

Organizational Agility: Autonomic Sensing and Adapting to Change

Leading organizations are motivated to enhance their business agility by increasing their ability to sense changes and greater capacity to respond and adapt to them accurately, quickly and confidently.  These organizations develop an autonomic computing framework of self-managing distributed IT resources by integrating a rich and flexible set of dynamically reconfigurable automated subsystems that are interacting with each other.

DRYiCE is HCL’s autonomics and orchestration suite of services, products and platforms powered by AI technology. It is a suite of built-in core technologies and IT automation service modules and interfaces to manage process interdependencies and workflows that realize sense-and-response monitoring components and intelligent visualization dashboards. IT organizations use DRYiCE to build advanced systems for further discovery, optimization and prediction on top of this foundation, thereby reducing cost and risk and accelerating time to value of cognitive autonomic solutions.

Again, not all business systems justify the complexity and skills of advanced machine leaning and artificial intelligence. But organizations should be encouraged to experiment with AI technology to explore its potential and understand the application development and data lifecycle of machine-learning systems. A platform that provides key technology building blocks is an effective low-cost approach to experiment, develop proofs of concept, and explore different solution roadmaps.


This blog article was sponsored by HCL Technologies.