AI can help organisations anticipate employee turnover and significantly improve retention rates, write UNSW Business School's Dr Andrew Dhaenens, Professor Mary-Anne Williams and Professor Karin Sanders.

Employee turnover is a significant concern for organisations globally. Retaining top talent is a critical component of business innovation and competitiveness. As companies invest substantial time and resources in the recruitment and training of employees, employee turnover is costly and disruptive.

According to the Work Institute (2022 Retention Report), the average cost to replace an employee is 33 per cent of their base pay. A recent PwC report found that 38 per cent of Australian workers plan to leave their job over the next 12 months, while the Australian Bureau of Statistics found that almost a third (31 per cent) of organisations find it difficult to retain staff.

Predicting employee retention has become a focal point not just for HR departments but business leaders and the entire C-Suite as well. There are several challenges that make predicting if and when an employee will leave difficult.

Identifying the factors that influence retention

One of the main challenges in predicting employee retention is identifying the factors that impact an individual's decision to stay or leave an organisation. These factors are complex and can include elements such as job satisfaction, work-life balance, compensation, co-worker/supervisor relationships (and difficult relationships with direct managers in particular), organisational culture, and career advancement opportunities. Understanding the relative importance of these factors for each employee in a specific business is a complex task, as individuals place varying emphasis on several aspects of their work experience.

Uncovering hidden patterns

Predicting employee retention often involves analysing complex datasets to uncover hidden patterns and relationships. HR professionals must sift through data related to employee demographics, job history, performance metrics, and engagement scores, among others. The sheer volume and complexity of data can be overwhelming, and the relationships between these variables are not always visible or apparent, making it difficult to draw actionable insights.

Accounting for external factors

External factors can also significantly impact employee retention, adding another layer of complexity to the prediction process. These factors can include economic trends, industry-specific events, a change of jobs for a partner, and changes in the competitive landscape. While these factors are beyond an organisation's control, they can significantly influence an employee's decision to stay or leave a specific company. Accounting for external factors in retention models can be challenging, as they require broader market knowledge and an understanding of macroeconomic trends.

The role of individual choice

Predicting employee retention is that the decision to stay or leave a company lies with the individual employee. Human behaviour is inherently unpredictable, and even with the most sophisticated models and comprehensive datasets, it is hard to predict an employee’s choice every single time. Accepting a degree of uncertainty is an essential aspect of any retention prediction model.

However, advancements in artificial intelligence (AI) technology have opened new possibilities for predicting employee retention. Machine learning algorithms can analyse large datasets, identify patterns, and make predictions with greater accuracy than traditional statistical methods. However, until recently implementing these advanced tools required a significant investment in both technology and AI talent. In addition, there are concerns around the ethical and legal use of AI algorithms.

Enhancing employee retention with AI technology

We have started to work with Toustone, an innovative Australian company, to evaluate a new AI system called RetainTalent they developed to predict when employees are likely to leave enabling organisations to take proactive steps and design targeted interventions to retain them. Toustone developed RetainTalent in-house. It is a unique Australian product using cutting-edge AI technology and becoming a leader in a rapidly changing market.

In a collaboration between UNSW Business School's Business AI Research Lab and Hybrid Work Leadership Research Lab, we investigated RetainTalent’s user interface, predictive capabilities, and the ethical framework Toustone used to develop it.

RetainTalent uses machine learning to predict if an employee is likely to leave an organisation, to help reduce the cost and negative impact of staff turnover. It can be used to identify why people stay and when someone is at risk of leaving, by determining the key factors that lead to employee’s deciding to leave. This enables organisations to design and implement more effective strategies and processes that increase retention rates, performance, and profitability.

To assess RetainTalent’s performance we looked at specific applications where Toustone developed a solution using clients' data to achieve an impressive 90 per cent-plus prediction accuracy. RetainTalent ingests data from a wide range of sources such as payroll, HR and demographics.

RetainTalent is an intelligent tool that provides a dashboard with high-level summaries and visualisations.
RetainTalent is an intelligent tool that provides a dashboard with high-level summaries and visualisations. Photo: supplied
RetainTalent ingests data from a wide range of sources, and this data is used to build predictive models that provide business insights to help organisations improve retention. Photo: supplied

Machine learning algorithms use this data to build predictive models that can be used to develop business insights that organisations can explore and utilise to design strategies that improve retention.

RetainTalent is an intelligent tool that provides a dashboard with high-level summaries and visualisations. HR professionals and other decision-makers can drill down on this summary data to discover underlying patterns, insights, and predictions.

Toustone adheres to a comprehensive policy that governs data collection and storage, and AI model development and deployment. This policy addresses important aspects of data collection and AI model development and deployment including fairness, transparency, accountability, privacy, safety and security.

Predicting employee retention is a difficult and complex business challenge with numerous factors and potential ethical risks and pitfalls at play. By leveraging ethical and responsible AI technology, organisations can anticipate employee turnover and take proactive steps to improve retention rates to significantly lower costs and boost staff engagement and productivity.

Further notes:

Andrew Dhaenens is a Lecturer in the School of Management and Governance, Mary-Anne Williams is the Michael J Crouch Chair for Innovation and Professor in the School of Management and Governance, and Karin Sanders is Senior Deputy Dean (Research & Enterprise) and a Professor in the School of Management and Governance at UNSW Business School.

The UNSW Business AI Lab is led by Professor Williams and works with its partners to discover new business opportunities, overcome challenges to AI adoption and accelerate the next generation of leaders and entrepreneurs. The Hybrid Work Leadership Lab is led by Dr Dhaenens and Professor Sanders to help organisations better employee outcomes via hybrid and flexible work strategies and arrangements.

For more information please contact Dr Dhaenens or Professor Williams directly.