Machine Learning – the Next Big Thing for ITSM in 2017

As 2016 ends and we look forward to another year in IT service management (ITSM), one wonders what we (ITSM pros) should be focused on in the next twelve months. There’s been a lot of buzz this year about things such as  DevOps, enterprise service management, customer experience and consumerization, and digital transformation. But I think there’s a wealth of opportunities for us and the businesses we serve in another area – automation.

Not the process automation that we’ve benefited from since the early days of ITSM tools, or the orchestration that has made virtualization and cloud so much easier. I’m instead referring to a different type of automation, where we “cede power to the machines” and their ability to learn, i.e. machine learning – “the study and construction of algorithms that can learn from and make predictions on data” (source: Wikipedia), where:

“Advanced machine learning algorithms are composed of many technologies (such as deep learning, neural networks and natural-language processing), used in unsupervised and supervised learning, that operate guided by lessons from existing information.” Source: Gartner IT Glossary

And Gartner recently stated that:

“Smart machines will enter mainstream adoption by 2021, with 30 percent adoption by large companies, according to Gartner, Inc. Technologies including cognitive computing, artificial intelligence (AI), intelligent automation, machine learning and deep learning fall under the umbrella term for smart machines.)”

But I’m far more bullish about machine learning from an ITSM and IT support (or for any service and support scenario) perspective, as many opportunities, and possible solutions, to use machine learning in our everyday activities already exist. And I think we will quickly see ITSM tool vendors partnering with machine learning technology partners to deliver against them.

ITSM Is Full of Opportunities for Machine Learning

ITSM pros have spent much of the last two decades optimizing IT service delivery and support in the enablement of business operations. Different operational elements have been addressed – from the adoption of best practice processes, the recruitment and training of the “right kind of people,” to the exploitation of technology (in particular workflow and automation, knowledge management, remote control, self-service, and more recently business intelligence). Much of this has been done to improve efficiency and effectiveness – it’s what we seem to do in ITSM.

But we still often place too much reliance on human effort, and human intellect, when we could cede the power – okay, some specific tasks – to the machines. For example, tasks where algorithms can be used to understand patterns and context to decide the best course of action without human input. The results and benefits being similar to the use of our existing orchestration-type automation:

• Greater speed/efficiency – machines can be quicker than the smartest of humans. They also work 24/7 including all public holidays.
• Reduced costs – people costs are still a large part of the overall IT department budget and, while technology isn’t necessarily cheap, the cost of automation should be more than covered by people-cost savings.
• Better people utilization – it’s as simple as freeing up time-poor staff from repetitive (and potentially mundane) tasks to allow them to focus on the more important things.
• Reduction in human error – people make mistakes and it doesn’t matter if they are inexperienced, rushing, or tired – these mistakes might have an adverse business impact. Automation makes far fewer mistakes, with these usually still people related.
• Greater ability to change – quickly changing ways of working, or even just responses to simple questions, can be problematic for people as they unconsciously flit between old and new for a while. Automation, on the other hand, just stops the old way and starts the new way.
• Better customer experience – while automation is often seen as a boon for the service provider, it also extends its benefits to the service receivers – whether it be speedier delivery, the passing on of cost reductions, greater probability of service delivery, or better support and customer service when things do go wrong.

A new AXELOS survey and report, “The ITSM Professional in 2030: A future full of opportunities” (January 2017), also shows that most ITSM professionals are already betting on automation and machine learning:

• Automation – 89% of respondents “think that an increase in automation will take over the repetitive tasks of IT, creating more time for service managers to focus on delivering more value to their organizations.”
• AI/machine learning – 77% of respondents “said they believed these trends would have a profound impact on the IT workforce, liberating ITSM professionals from routine tasks and free up time for responding to demands for more creativity and ‘human’ input.”

So where can machine learning help?

Five Examples of Machine Learning Use Cases in ITSM

These are all things that can be done or used now:

1. Identifying and predicting issues and problems. The technology can also offer up the most likely resolutions. It will reduce mean time to resolution (MTTR) and opens the door for predictive maintenance (and less reactive fixing).
2. Better understanding the risks of proposed changes. Not only understanding what will be impacted but what the likely impact will be.
3. Predicting what’s needed or even what will happen. From demand and capacity planning to understanding the future levels of customer satisfaction based on what’s happening or planned.
4. Greater access to information and known solutions. Machine learning improves search accuracy, with data-based recommendation capabilities similar to what people already get in their personal lives from companies such as Amazon and Netflix.
5. Easier and better knowledge management. The last two decades have shown that people aren’t great at knowledge management – or the knowledge article creation part of it at least. Machine learning can be employed to both identify “missing” articles gaps and to create new articles automatically from existing tickets, i.e. the already documented resolutions.

So what do you think of the possibility of machine learning becoming a staple of ITSM? Or are you already using it? I’d love to get your feedback in the comments section below.