Author: Adrian Baillie-Stewart — Director and Content Strategist at Content Strategics (Pty) Ltd

Editor’s note — This is the fourth article in the four-part series of articles published exclusively for Paymaster People Solutions. The full series of articles looks at the growing prevalence of Artificial Intelligence (AI) and its impact on the Human Resources (HR) industry. This—the final article in the series—explores AI within the context of payrolls and payroll functions. For your reading pleasure, the first in the series is available, here, the second here, and the third article here.

Introduction

The four-part article-series commenced by considering what Artificial Intelligence (AI) is, and why AI is currently becoming increasingly more relevant to the operational running of the human resources (HR) function of an organisation.

Upon defining AI in the first article of this series in AI, it seemed appropriate that, for an industry that is commonly known as “HR”, there ought to be a keen interest in emerging technologies that best-leverages human-like thinking for practical applications within the various specialised fields of the human resources function. These include recruitment department, employee training and development, and payroll.

Andrew Anderson, writing for ComputerWeekly.com’s MicroScope, states that “AI is only materially useful when it can mimic what makes the human brain so unique – the ability to learn, change behaviour, and make decisions based on new rules and apply them to varied and unstructured data.”^[1]

Consequently, let us consider AI within the context of a payroll department and payroll function of an organisation.

Increased expectation for HR to adopt new software technologies

Most HR professionals know that there are increasing expectations for the human resource function within an organisation to adapt to numerous technological advancements. For this reason, the software development trend is evolving towards human capital management (HCM) software applications that are increasingly beginning to have AI functionality programmed into its core processing engine. In some cases, AI software capabilities include the ability to engage in some self-learning of its own. This self-learning ability (embedded in software) is more commonly known as deep learning. But first, let us turn our attention to a short discussion on AI for payroll.

AI for payroll

We start with a question: are you considering the procurement of an AI-embedded software solution for the first time? It may be interesting to know that, for most organisations who still make use of manual processing methods, or outdated legacy payroll software, it’s the payroll department that best benefits from an AI-enriched software solution.

Brian Radin states that “payroll contains perhaps the greatest amount of administrative activity of any human resource process, with every pay cycle demanding accuracy, timeliness, and well-planned coordination across multiple organizational entities. The demands come from both inside and outside the business.”^[2]

Within the HR industry, organisational efficiency of its people management remains paramount. For this to be done, HR needs to understand their workforce. HR needs to make sure that their workforce’s payroll is always processed on time and that upon completion of a payroll-run, that it is always 100% correct. This is made far easier when AI-enriched software has a footprint in all of the HR functions: the recruitment department, the employee training and development department—and of course—the ever-important payroll department.

With AI-enriched software, management decision-making is also greatly enhanced by giving real-time access to data-drive statistics and reports that can be easily accessed by a variety of employee levels within the organisation. Access to these reporting-resources facilitates the outline of a far clearer picture of the workforce (at any given point in time), particularly with regards to the payroll function, where problems need to be identified early so as to prevent crisis management. For AI-enriched payroll software to be truly helpful to the organisation, some form of ‘self-learning’ is required of the software — it’s called ‘deep learning’.

Deep learning stems from artificial intelligence

Writing about deep learning^[3], Luis de Abreu—a software engineer specialising in AI software development—makes reference to payrolls and the payroll function to illustrate a critical point about deep learning which stems from AI-enriched software: namely, that humans are well-suited to assessing the accuracy of a payslip in terms of its content — not for whether the calculations are correct, but whether the payslip appears (according to human logic processes) to be correct, or whether the salary-scale appears to be correct, or whether the employee age-criteria appears to be setup correctly.

Traditionally, the detection of non-obvious inconsistencies with data fields on a payslip were flagged by experienced payroll administrators with a keen eye for recognising these non-obvious errors. With AI-embedded software however, this is all (rapidly) changing. Now, deep learning accomplished by a computer, gradually facilitates the detection of likely non-obvious payslip errors, thereby making AI-embedded payroll software so much more efficient at processing 100% accurate payroll runs without any margin of error — realisable over a period of time.

This type of ‘AI learning’ which the computer is (seemingly) able to ‘learn of its own’, is called ‘efficiency-by-experience’ learning which stems from AI-embedded learning algorithms: the more user-data that the algorithm is exposed over a period of time, the better the human-like corrections and predictions to be outputted by the software.  An excellent example of deep learning in action, is to recall the way in which our mobile smartphones are (seemingly, of its own) able to learn our personal keyboard-usage preferences, as well as calendar scheduling phrases and names, or our regularly-used email addresses. This deep learning capability which our smartphones can perform, significantly reduces our user-involvement in many repetitive inputting-sequences within various apps and when navigating the operating system itself. However, notably so, it is only over a prolonged period of time, that—as a smartphone user—we begin to recognise, for example, how the predictive-text functionality on our phone improves with time — almost as if the phone has learned what the user’s phrase preferences and word selections are. This is exactly what deep-learning entails. Isn’t it amazing that artificial intelligence systems can identify patterns and create connections that would otherwise be too laborious and time-consuming for humans to enjoyably repeat thousands of times.

Concluding thoughts

The contemporary status of the HR function is such that human resource management teams are expected to assume significant accountability for some of the organisation’s overall key performance indicators (KPI’s) that ultimately stem from within the HR function — KPI’s that lead to the overall profitability of the organisation. Leveraging AI-enriched software for deployment in a variety of HR settings is probably the most prudent strategic business-decision that could be made for any progressively oriented organisation (particularly in the mid- to long-term).

At Paymaster People Solutions, its progressive and forward-thinking team of HR professionals have already assumed these significant accountabilities and KPI’s. Consequently, Paymaster People Solutions employees are now proudly helping their clients to gradually transition towards the adoption of these changes with much excitement too.

———END———

Sources

^[1] Andrew Anderson, ‘Time for Artificial Intelligence to Make Life Easier’, MicroscopeUK, accessed 24 February 2017, http://www.computerweekly.com/microscope/opinion/Artificial-Intelligence-that-wants-to-make-your-paperwork-easier-not-throw-you-out-of-the-airlock.

^[2] Brian Radin, ‘How Robotic Process Automation (RPA) Is Shaping the Future of Payroll’, accessed 24 February 2017, https://www.cloudpay.net/resources/how-robotic-process-automation-rpa-is-shaping-the-future-of-payroll.

^[3] Luis Gomes de Abreu, ‘Artificial Intelligence for Payroll and HR Applications’, LinkedIn Pulse, 11 July 2016, https://www.linkedin.com/pulse/artificial-intelligence-payroll-hr-applications-luis-gomes-de-abreu.

Paymaster People Solutions (Pty) Ltd is a subsidiary of Grant Thornton Advisory Services Cape (Pty) Ltd
Grant Thornton Advisory Services Cape (Pty) Ltd is a member firm of Grant Thornton International Ltd

© 2016 Paymaster – Online Payroll Solutions | All Rights Reserved