Robots Stole My HR Job

It was our privilege to host our inaugural annual HR Tech Korero series with Drew Williams and the University of Sydney team. 

Sydney Uni kindly introduced us to their cutting edge HR bots, of which there are over 300 RPA (Robotic process automation) bots scattered throughout the University, and a dedicated AI team supporting their implementation and maintenance. 

The Sydney HR operations team showcased four of their most powerful and sophisticated bots  covering requisition creation, onboarding new employees into superannuation and ServiceNow case attribution. For those that are new to bots, RPA Bots are a way of programmatically creating process steps that can add, edit, change, attach, delete any information in any system at scale, speed and are relatively inexpensive…. well compared to annual HCM Tech license costs.

Before we got started on the nerdy stuff, Drew and team kindly set the scene first by outlining the huge two year journey implementing Workday and Servicenow, and the volumes they deal with which I will be honest far our strip any thing we have seen in NZ (happy to be corrected otherwise)  i.e. 110,000 HR Service Now tickets a year and 90 hires a day! 

Their Bot journey started with a process automation lens as they were simply buried under the volume of transactions.  They noted that to make the bots pay for themselves, and actually make a difference justifying the expense, high volume processes were targeted. It didn’t matter what system(s) was used, RPA bots could mimic the logic, process steps, data entry of a logical human and could apply machine learning to improve on its own results. Pretty impressive! 

They demonstrated one of their first bots which was triaging ServiceNow cases and attributing them to the correct resolver groups. Pre RPA, they had an approx 30% reject rate where the assignment wasn’t correct, post the implementation of their Bot this reject rate was down to 6% and improving through machine learning! For me the process efficiency wasn’t the true takeaway, it was more the unattended consequences! 

The Korero attendees quizzed the Sydney team on the obvious “what happened to that capacity release?” and yes…we all fully expected a net FTE change – tick in the box for robots and accountants but the surprising outcome was actually the opposite.  Sydney realized that the way that they had been triaging and automatically closing cases (link to knowledge article and cue spin/cycle of unhappiness) while fast and efficient was a terrible user experience. So the additional capacity was moved to real life human interaction and helping people on the phones – Hi my name is x and I am not a Robot! This is a real step change and ended up improving their customer satisfaction stats which was traceable directly to the introduction of the bots.

One of the other bots demonstrated for the payroll people reading this was collection and categorisation of superannuation information of new starters into Sydney’s payroll system zzzzzz. Ok not the most exciting juiced up bot user case but a really good example of how bots are not just limited to routine tasks but can handle subtle complexity and sophistication. This particular bot was set up to extract superannuation option details from an external site, collate it and then based on business logic and calculations enter this information directly into the payroll system – every day, like magic (no capacity issues, no fatigue, no personality….(for now). 

The complexity is up there, and to underpin this, two of the world leading HR systems could not collect/ automate this process without a $ignificant amount of investment and config.  

Once again…… the end result wasn’t about signing the leaving cards for the various payroll teams so they could re-focus on payroll audits, reconciliations, the stuff which matters, paying your people correctly the unsung, unrecognized, underappreciated function until…… 

Ok I have to honest even for a self confessed geek, viewing a bot in action is quite boring, the epileptic inducing flicking of dozens of screens and data, which I think was purposely slowed down so we could visualize the process, was like watching paint dry, but the hidden message or the power of this demonstration was the business rules, logic and sophistication which was required to make this work was soooo underplayed.  Snippets of how much work was involved bubbled out throughout the show and tell. The journey time was significant, training a bot takes time and investment and it isn’t a finished job, it needed constant feeding and watering as the bot is so reliant on the the structure and UI of the subsidiary systems. Finally there is the constant quest for perfection and amending the business rules/logic to obtain that near perfection. I will say though, Sydney Uni were quick to point out that Bots for them were never thought of as the final solution and only another tool to support their quest for HR process efficiency which I think is a position they came to after going through the significant investment. 

I do think this is the future even at a smaller scale as this gains industry momentum , barriers of entry reduce and HR bots become commonplace. It really does have the ability to truly truly change the HR tech and HR function landscape. 

RPA bots won’t steal your job but they will definitely change it! 

Acknowledgements to: