These days it’s not uncommon to open a newspaper or news app and find dire warnings about how humanity is dangerously ill-prepared for the threat posed to certain jobs by advances in various forms of Artificial Intelligence and machine learning, collectively referred to as Automation. As someone who has worked with automated timetabling for many years, the differences between 'timetable automation' and these emerging Automation technologies (like self-driving cars and other AI) are very clear, but I have realised that it would be all too easy for others to conflate the two, leading to some unrealistic expectations and fears.

This blog post is to address the possible fears first: That CELCAT Automation can take the place of human timetablers, leading to job losses. This was recently mentioned to me by a client at the start of an Automation implementation project, and I realised that she had been seriously worried about losing her job, purely because she did not fully understand what CELCAT Automation does. The truth is that Automation is so unlikely to replace a human timetabler that I almost feel more than comfortable to say ‘never’.

While it is true that Automation can speed up certain repetitive tasks, in my experience these types of tasks never make up the bulk of a timetabler’s responsibilities, even though they could take up a lot of time during certain parts of the year. Instead, timetablers must be able to come up with creative solutions to unexpected problems, use local knowledge of resources and people to make complex decisions, and deal with exceptions to established rules and patterns, often under incredible stress and with very little time.

This is not what CELCAT Automation does. It won’t suggest that you quickly phone up Professor Smith who owes you a favour to ask if she’d be willing to move her class to the building next door on short notice so that you can fit in another group of students you just had to relocate because the projector in their classroom stopped working. Instead, CELCAT Automation can free up a few days or hours of your time during the timetable construction phase of your annual working plan, allowing you to focus on other, more complex tasks that require the human touch, so to speak.

It could, for instance, do eighty to ninety percent of your straightforward room assignments in a few minutes, leaving you more time to deal with the remaining complex ones that require careful negotiations or relaxation of ownership rules. If you’re lucky and local resources allow it, it could do one hundred percent of your room and/or time and/or room assignments, but to do that, it would have to be set up first.

This brings us to the other aspect of this: Automation needs a highly skilled human operator. Someone must set up and maintain the data, rules, constraints and their relative weights. Someone must make sure the rules reflect the policies of the institution, that changes are incorporated into the model, and that the output from the Automation is acceptable for end-users - students and staff. The ability to set up the rules and interpret the output is something that only an experienced timetabler will be able to do, and as such, Automation is to be seen as a tool and extension of the timetabler and his or her abilities, not as a replacement.