The former is usually managed by the infrastructure and operations (I&O) teams, whereas the latter is done by the application development and delivery (AD&D) function. As 70 percent of IT budget goes towards keeping the lights on (KTLO), artificial intelligence for IT operations (AIOps) began as a way to drive cost reduction through greater automation of operations and infrastructure activities, shifting the burden of KTLO from humans to machines – the ultimate goal for AIOps.
However, for modern day organisations, applying AIOps for I&O teams alone is not enough. For organisations to reap all the benefits of AIOps, it needs to be part of the application development organisation’s strategy to reinvent the business. Enter Value Stream Management, or VSM – an approach which encompasses every step in the software delivery process and which is centred around helping companies move faster with higher quality, whilst helping them to align technology activities with business outcomes.
Incorporating AIOps into VSM provides all streams of the business with data-driven insights that help organisations to simultaneously manage risk and deliver high quality digital services more efficiently.
The development of and key roles of AIOps
Recently AIOps has been applied to many use cases. The first wave involved the use of AIOps for event noise suppression by filtering out unnecessary alerts generated by application, infrastructure, and network monitoring tools. AIOps techniques then extended to automating the process of understanding the root cause of issues to enable the swifter remediation of major incidents thus helping to reduce the downtime of critical business services.
AIOps has also found applicability in the area of service management. AIOps techniques have led to the development of tools such as chatbots to help provide an answer to an employee’s query by leveraging the company’s knowledge base system as well as historical incident and service request patterns.
Another compelling AIOps use case in this area is the application of machine learning techniques to predict whether an incident that seems relatively benign right now has the propensity to become a major incident based on patterns detected from historical IT service management (ITSM) and monitoring data. Additionally, AIOps and process mining techniques can help uncover bottlenecks in service delivery such as recurring patterns of ticket reassignments, clusters of incidents that may have similar underlying characteristics even though they are not tagged as such, etc. These insights can help drive process optimisation and automation decisions improving the cost, quality and effectiveness of service delivery.
As AIOps matures as a field, we are seeing more extensive capabilities such as change risk prediction and release schedule risk prediction come to the fore, highlighting the growing applicability of these techniques to the world of DevOps. With the broadening of the relevance of AIOps beyond the initial shores of the IT Ops organisation, it also has a key role to play in automated governance across the IT estate, facilitating the transition from a realm of key performance indicators to one of machine generated key risk indicators. The benefits of such a transition from lagging to leading indicators are extensive and range from lowering costs of IT operations and providing a higher quality of service, to maximising uptime of services and ultimately lowering risks for the business.
So how does AIOps benefit Value Stream Management?
There is currently a mounting dissatisfaction with the ROI being realised by digital transformation initiatives. In fact, a recent survey of 600 Business, IT and Security decision makers found that the majority of enterprise leaders are concerned about just that. Businesses are learning that practices such as Agile and DevOps are necessary, but not sufficient. To reap the full benefits of digital transformation implementations and optimise business outcomes, organisations need to address the challenge of how best to tie the work of the development team with the needs of the business.
Additionally, as incumbents embark on their digital transformation journeys, they must be able to release new capabilities rapidly without impacting the availability and performance of the production systems that power the business. How to move faster with quality is another key challenge.
It’s no small task but, when implemented properly, VSM can help address these two challenges and be the missing piece required for digital transformation success. The required capabilities in this area have been well described in recent research by Gartner and Forrester. Their research shows that one fundamental impediment is that most enterprises use multiple point products in their development and delivery toolchains. This siloed landscape limits the ability of organisations to harness data across development, operations, and the business to support the goal of moving faster with quality. Gartner states that the solution for this is a VSM platform (VSMP) “that enables business and IT to align their goals,” and helps “organisations to use leading indicators and predictive analytics to gain a competitive advantage and quickly adjust course to drive customer-centric outcomes.”
As one may have already surmised from this description, AIOps techniques will have an increasingly important role to play in VSM. As an example, one of the required capabilities for a VSMP highlighted by Gartner is change risk analytics, a capability also described in their AIOps market guides. A change can be as small as a bug fix or a modification of a configuration, or as large as an entire release. Change is where the work of development meets the world of operations. It is also the largest cause of outages and business service interruptions. Change risk prediction uses AI techniques to analyse historical change data from across the software development and delivery lifecycle to understand the drivers of failed changes, quantify the impact they have on business services, and score the risk of upcoming changes failing. This allows an organisation to automate the process of pushing low risk changes into production and also hit the pause button to further scrutinise a proposed change deemed risky by the ML model before it goes into production. The ability to do this is at the very heart of being able to move faster with quality.
Similarly, there are other AIOps use cases such as release schedule and quality risk prediction, as well as automated governance across development and operations that are fast emerging as foundational capabilities in the VSM approaches of companies seeking to optimise and transform their businesses through digital transformation.
AI and humans working in harmony
AI today has great applicability in both decision automation and decision augmentation. As consumers we experience decision automation via new tools like automated appointment schedulers, whereas decision augmentation capabilities are regularly woven into our online shopping experiences such as relevant recommendations for books to read or items to purchase based on shopping history and inferred preferences.
The same applies to IT organisations though, as we saw in the AIOps for VSM examples just discussed, the focus is largely on decision augmentation. This is inherently a coexistence model where the goal of the AI capabilities is providing machine generated insights into monitored patterns and delivering early warning of risks to arm individuals in specific roles with the proactive insights they need to fulfil their responsibilities efficiently and effectively.
Looking to the future
With businesses expanding and systems evolving every day, the challenge of delivering higher value at lower costs gets tougher. AIOps, as a way of shifting some of the growing burden from humans to machines, can help ensure business continuity and drive efficiencies across the different departments.
Companies that choose not to consider the AIOps-powered VSM approach are missing a significant opportunity. VSM is all about smart decision making, being able to predict risk and avoid a drop in operating efficiency. As the world continues to invest in digitisation, businesses that embrace all AIOps has to offer will have a lasting competitive advantage.