The Role of AI in HR Decision Making
An AI-powered world
“Our AI-powered tool will help you make data-driven decisions. Can we call to see how we can help you?” - reads the umpteenth email in the HR manager’s inbox from various HR vendors.
When it comes to AI-powered insights, there are many concerns floating around regarding data privacy, data completeness, and of course, matters of integrity and welfare. On the other hand, every department seems to be using AI and data to solve their problems: marketing, finance, product teams, the list goes on… so, why shouldn’t HR follow the trend?
AI has had a profound impact across many functions, for example: marketing now uses predictive analytics to automate campaigns without human intervention so that it can act 24/7; and finance applies AI in their investments, expenses, and case evaluations. Of course, HR is also already successfully using AI, especially in Talent Acquisition and L&D through recommendations for both employees and employers, as well as in the budding areas of People Analytics and Workforce Planning. And yet, despite early successes and headline-grabbing use cases in HR, there are specific challenges facing HR which has led many to think, “ “perhaps HR just isn’t mature enough yet”...
Googling “HR maturity model” retrieves over 20 million results from across the internet, many of which discuss the strategic and business-integrated future of the function. Almost entirely, the upper echelons of these models can be reached after adopting cross-functional and AI-powered predictive analytics, assuming vast collections of data are available for HR to work with. However, according to research from KPMG, “58% of CEOs still see HR as an administrator rather than a value driver”. What we see here is an ‘impact gap’ - between what HR is capable of and the value it is currently perceived to deliver for the business.
Let’s take a closer look at some unique challenges that HR must overcome in order to deliver value with AI-powered decision making.
Challenge one: Humans are complex, resources are complex
By its nature, technology thrives in well-defined environments, be that the data, the components, or the decisions to be made. Unfortunately, humans – the very H in HR – are notoriously complex and hard to define.
Let’s dive a little deeper. What does it mean for an environment to be well-defined? On a high level, it translates to ‘unambiguous’ and ‘with a unique interpretation’. For an environment to be well-defined, we require, among others, complete visibility of all the components and their boundaries, an understanding of the behaviour of these components, and some notion of ‘what’s next’.
A brief thought experiment to engage with. Visualise your organisation in some way – maybe as a graph or as a diagram, for example. Next, draw a circle around what, for you, is ‘Engineering’. And then, draw a circle around what you consider to be HR. My guess is, the first circle was rather straightforward to draw, whilst the latter may have raised a few different ideas.
Within the organisation, HR could be the collection of the processes and employees that work directly for the HR department, but it could also cover every single employee within the organisation as they are the human resources. Reflecting on what a well-defined environment is, this brief thought experiment would suggest that a lack of consensus around the boundaries of HR may be an apparent challenge in HR’s work.
Understanding the components, aka the employees, would be the next dilemma. People aren’t a problem to solve unlike a product, for example. Next to being the link between company-wide objectives and people strategy, HR has the responsibility to lead, support, and educate employees in all functions. These elements are unique for each individual in the organisation, with many variables from social and personal experiences as well as trends. Understanding each employee’s personal values, ambitions, and behaviour is indeed quite the complex task.
Challenge two: HR serves two masters
Arguably, the most important success criteria of all functions is to satisfy their stakeholders. It’s also true that the more stakeholders there are to manage, the more complex the governance becomes. For HR, these stakeholders are the Business on one hand, and the workforce on the other.
The requirements of Business could be the following: the right people at the right place at the right time, ensuring compliance with regulations and external bodies and achieving all of this within a budget. The requirements of the workforce could be to: feel respected, acknowledged, provided with growth and opportunities, and to have security and trust in their environment.
All of these requirements should be considered when making decisions within HR. While there is often overlap in potential outcomes, these interests will sometimes collide. Therefore, we could define this second challenge as ‘HR serves two masters’. Such a split contributes to the complexity of decision making in the workplace; in each direction, HR is fighting an uphill battle of eternal appeasement.
The unique combination of people and data-driven decision-making in HR
The table below illustrates exactly how big the difference between HR and other departments is. From their subject matter to their client, HR really holds a unique position within a company.
| HR | Other Functions | |
|---|---|---|
| Analysis | People (ill-defined) | Product (well-defined) |
| Predict | Individual behaviour (individual, unique) | Market (aggregate, mob/group movement) |
| Stability | Volatile (highly dependent on trends in technology, market, society, and individual integrities) | Convergent (product-market fit, dependent on trends in technology, market) |
| Client | Business, Employees | Business, Customer |
| Metrics | Headcount, hours, skills, as well as money | Money |
How can HR use AI to aid decision making?
Moving forward, as with all other functions, it will become increasingly necessary to adopt data practices to support decisions and processes in HR. HR will continue to be pushed and pulled by the business and the workforce. In this delicate balancing act, HR must serve as the interface between the objective (data) and the subjective (people) sides of decision making.
As such, being aware of the situation and the complexity is but one chapter of this story, so here we offer a take on the next: data-augmented decision making. Automating decisions with AI and data can be overly rational and maybe incomplete for tackling the challenges of HR, which is why what we really should be talking about is augmentation. Where does AI and data help in the HR decision-making process? To answer this question, we offer a model to consider:
Let’s talk through an example of reskilling and upskilling – one of today's hot topics. Company X is in an industry in full digital transformation. They have identified that they will require more people with digital skills if the company, as a whole, wants to remain competitive. Commencing with the business question, they start with ‘could’:
Could we fill these positions leveraging the resources and skills we already have?
In this approach, they look at their employee data (such as skills, location, compensation, etc.) to determine availability and costs; luckily, the data returns a large body of employees who are now working in the retail division of the company. Their customer-facing skills could be used in online customer service.
The high-level profiles match, so the CHRO gives the go-ahead to move forward.
Unpacking this decision to a personal level, HR then looks at the individual cases.
Amongst them is Jo, currently working in a physical store location:
Should we transfer and reskill Joe?
Should we backfill Joe’s current role? Does a transfer leave a critical hole behind?
They look at the data again as the AI spits out the answer: Joe has quite a few skills that are adjacent to the skills the company needs. The company is planning to close the retail location where Joe works in 3 months, so there is no need to backfill Joe’s role.
“Problem solved! Let’s get Joe over to the growing e-commerce department!”
Not so fast! There are other aspects to consider. And this is when the decision-making process in HR has to become more subjective and depend more on the individual’s perspective – not just what the data dictates. Does Joe want to move to e-commerce? Does he fit the culture in the team? And so, a final set of questions must be asked:
Would Joe want to move to this role and go through the training?
Would Joe be a good fit personality-wise?
This, of course, isn’t about reinventing the wheel, it’s about understanding the role that AI-powered data analytics plays, versus the individuals impacted by the decision. The story becomes infinitely more complicated when there are thousands of employees needing to be reskilled, and thus thousands of decisions are being made. Finding a way to speed up the process, and simultaneously respecting stakeholders, is the critical component in managing the complexity of a technology-backed HR function.
HR is a curator of the objective and a guardian of the subjective. AI-powered decision-making is here to stay, so understanding where this technology excels – and where HR needs to depend on human-powered decision making - will be an essential skill in adapting to the future of work, to serve both the business and the employees.
As technology evolves so will HR.
ABOUT THE AUTHOR
Ben Searle is part of the Product team at TechWolf.ai, helping companies in transformation with an automated, continuous and objective view on the skills they have in-house, the skills they need and the gap in between. Ben is hooked on the future of HR and organisational psychology and has a background in computer science and psychology, with a specific interest in Human-Computer Interaction.
Caroline Styr is the Research Director at Insight222. She is a thought-leader, researcher and writer on people analytics and the future of HR. Prior to joining Insight222, she worked at the Center for the Future of Work where she was an advisor and in-demand speaker on topics related to the future of work. She has also held roles in digital services and transformation consulting at Cognizant. Contact Caroline at caroline.styr@insight222.com
Build the HR Skills of the Future with myHRfuture
The myHRfuture Academy focuses on supporting HR professionals build the skills they need to succeed in the future. It aims to help you become more data driven, business focused, and experience led – to acquire the skills necessary to process, produce and leverage digital information. Our new HR skills of the future assessment not only shows you how you stack up against the 9 skills of the future HR professional but also highlights your development opportunities, providing you with a personalised learning journey that aims to guide you every step of the way so that you can close your skill gaps, deepen your knowledge and Press PLAY on your Career.