Operators are starting to accelerate their journey to automation of their fault management processes. However, there is a long and tortuous journey for most to achieve the fully autonomous nirvana envisioned for operations. There are too many mistakes being made in planning the journey and too much manual activity not included in the current plan which prevents any prospect of a transformation to a zero-touch future. Surveys show that 72% of automation initiatives fail to deliver business outcomes. There is much excitement in the industry around the capabilities of new and re-invented automation technologies from the past. The sad truth is that they are not working, and no amount of marketing injection of the magic sauce of “machine learning” and AI will make the technology automate more.
Why do companies fail to achieve success transforming Fault Management Operations?
The problem stems from the fact that these technologies propose to provide a one-hit “silver bullet” to the beleaguered automation programmes. The re-invented tools just look prettier and have easier user experiences than previously. However, they still fundamentally perform the same function as they did 30 years ago. With the same, unsuccessful, outcomes. This reinvention and recycling of technologies including ‘bots, and scripts in modern languages like python, and PowerShell; the re-branding of existing workflow, collaboration, case and task-scheduling tools as “automation”, “orchestration”, or “intelligent”; and the evolution of screen-scraping macro engines (reinvented as “Robotic Process Automation”), has not fundamentally moved organisations away from the same task-automation, engineers were creating a generation ago with simpler tools.
So how can organisations transform Fault Management with Automation?
The type of transformational automation required in fault management and remediation is a broader approach transferring work from a people-centric operation to a machine-centric one. It requires a plan which evolves through several automation styles and approaches. Cortex invests extensively in understanding and defining how to make this journey successful. This ensures that the evolution continues toward the autonomous nirvana whilst delivering value at every stage.
Functional automation is improving in complexity and richness especially in the new generation of network equipment, – EMS’s, OSS’s, SDN controllers, etc. It is easier to assist in better analysis and decisioning in a workflow, or ticket-based, environment. Tools such as RPA bring the opportunity to automate tasks using the existing screen and keyboard controls. However, RPA introduces challenges in the improvement and maintenance cycles, as well as the end-to-end performance. These kinds of tools will help in an assistive context where scale, velocity and capacity are not an issue. But beware – task automation rarely reduces headcount despite that being used as the primary justification for implementation.
Will Machine Learning and AI deliver the outcomes your organisation needs?
Machine learning and AI capabilities are causing a lot of hype but are currently offering little practical benefits in the fault management arena. By design their outputs are probability-based, and actions based on them could be wrong with potentially disastrous results. Even when the models are highly tuned, the dilemma between maintaining a static model with RPA (that will quickly become obsolete) and providing a dynamic model through machine learning (which will learn about all of yesterday’s problems) is a trap leading to failure. Both will eventually be self-destructive.
Intelligent Automation, however, leverage’s the power of machine intelligence in analytics and decisioning using deterministic techniques to drive rules, heuristics, and models for predictable outcomes. This will accelerate any automation initiative on the road to autonomous operations. Intelligent Automation introduces intelligent analysis, intelligent decisioning and intelligent cross-process orchestration capabilities. To drive both autonomous processes, and third-party task-based tools which will transform the approach to fault management. It’s only by introducing real process automation and then embedding intelligent analysis and decisioning that companies can release their expensive fault management engineers into more valuable roles. For example, new product development and complex customer solution sales and design.
Read the complete Evolution of Fault Management Whitepaper to understand how CSPs can bring together the strategy and methodologies to maximise successful outcomes.Read more about our successful Telecommunications Use Cases