Automation today, and certainly in the future, can do many things equally if not better than humans, providing efficiency and cost savings. Ideally this also allows people to do more interesting work.
The opportunity is clear that – as artificial intelligence becomes yet more intelligent – systems will be able to scan data quickly, make better observations and create more meaningful links to make decisions; in other words, automated systems will be able to understand a complex problem across multiple companies and systems in seconds, then take remedial action before anyone notices.
However, we have to be mindful of the hype that accompanies this too: why would you want to use this capability? What benefits are you expecting? Do you use it for its own sake?
To automate or to not automate
What makes sense to automate? For example, standard, repeatable processes like buying a parking ticket in a car park. The technology to do this has become much better so it can be automated relatively easily and is more acceptable to the user. For IT Service Management (ITSM) the obvious standard processes to automate would be password resetting and request management.
Conversely, the scope of responsibilities typically handled by a service desk and ITSM team are, by definition, much more complex involving a large amount of non-standard work. There might be thousands of potential scenarios where just to log an item could be hugely complicated.
The reality of automation is still relatively challenging and it’s easy to make requests that confuse most of the digital assistants available on the market, resulting in a substandard experience.
If we do still assume there will continue to be more automation of work in the near future, we also need to grasp a key point around readiness for automation – data quality. There’s no point automating a bad process, however it’s just as bad to automate a process that uses bad data.
Making automation work
- Cultural acceptance: if people are used to working in a certain way you have got to be sure they will accept a change to automation, or it’s not worth doing.
- Data quality: automation can be a blunt instrument because it will do only what you tell it to do. So, the quality of the data you’re using is essential – you can’t automate if data is flawed or incomplete.
- Configuration management and change management: traditionally, the Configuration Management Database has been an inventory of what an organization has and how things relate to each other. Today, the challenge is keeping an inventory up-to-date as the world becomes more complex. This is where you need good change management to keep things up-to-date and accurate.
- Knowledge management: while organizations are getting better at helping people to help themselves through self service, it is still a hot topic.
- ITIL® and getting your “house in order”: ITIL processes have become hugely more important in doing things properly and getting the basics right; in fact, they’re absolutely essential. In the past people have cherry picked from ITIL but now elements such as configuration management and knowledge management can’t be sidelined.
As the update to ITIL best practice will help clarify, these are the starting points for any organization to get the benefits of machine learning and big data and it must be part of a strategy linking business, IT, process and service management; in other words, you can’t just “buy a bot” and expect it to change your life.
Read more AXELOS Blog Posts from Barlay Rae
ITIL, DevOps and the search for silver bullets
ITIL Practitioner: measuring what matters in ITSM
Getting human to human in ITSM
Meaningful metrics in ITSM: it’s about the business (stupid)
Tackling the rise of bimodal IT and 'two-speed ITSM'
What should IT Leaders be focusing on in 2015?