When data scientists explain Machine Learning, the potential from a HR business standpoint seems endless. We could predict…
• Who’s at risk of leaving our company
• Who will be an outperformer
• Who are the best candidates for a vacant position
• What would be the right recipe (prescription) for increasing employee retention, lowering absence rates, etc.
The possibilities truly are endless. The models, algorithms, tools and computing power are there.
What about the data?
Data needs to be of good quality and abundantly available to generate relevant results. Having data on a population of 1,000 people over a 5-year period probably isn’t enough to train a Machine Learning model to come to significant conclusions. In our case, anyway… it wasn’t.
One issue is this: administrative HR data – mostly stored in state-of-the-art HR Information Systems – is often too limited to reach significant conclusions!
This problem could be overcome by enriching the centrally stored HR data with insights from social media, email usage, sentiment analysis, other external data… But this brings us to the limits of what is ethically justified. Especially in this new GDRP era.
So lots can already be done, but there are still some data hurdles on our way!
Consider this scenario; you’re ill. Normally you would go the doctor, get your sick note, come back home, open your PC, logon to the network using a VPN connection, open your HRIS application (if you can find it), go to the sickness registration page, enter start and expected end date of your sick leave and upload the note… Not something you look forward to when your nose is dripping and you have a 40°C fever.
But simple administrative transactions like this are perfect for chatbots. Your company’s chatbot can inform your manager, activate your ‘out-of-office’ and upload the sick note to the right place. These cases are already being implemented today by using common bot frameworks with some cognitive intelligence behind it.
It just gets a little trickier with more complex cases (like cutting your sick leave short because you’re feeling better…). In those cases, it’s still good to have a human ‘back-up’ to deal with the out-of-the-ordinary.
Virtual Reality has hit the consumer market; however, adoption is still below expectations. The reasons are obvious: the investment from both content provider and consumer are quite heavy and the results don’t yet deliver the type of experience users expect.
But the tide is turning.
The workplace has changed dramatically (anywhere, anytime, any device, anyone) and there is now serious traction regarding the use of VR for connecting coworkers.
In short, AI for HR holds great promise and much is already possible today, especially in the fields of chatbots and VR. But there is still a lot of AI ground to cover for HR people everywhere.
Want to learn more about AI and human resources? Check out our report (in Dutch):
“HR digitalisering in Vlaanderen – vandaag en morgen. Een huidige stand van zaken en vooruitblik op de toekomst”
This report, prepared by delaware and vacature.com, contains a section on AI in HR and is downloadable for free, online.