A recent study discovered that over one-third of Filipino employees are poised to depart from their current roles within the next year, presenting a persistent challenge of high turnover rates for HR departments and companies. HR frequently needs to develop effective hiring and retention strategies, hindered by the need for rapid data analysis, which can lead to misaligned expectations and reduced productivity. How can companies and hiring teams effectively navigate this challenge?
For my column at the Sunday Business & IT, I interviewed Jayant Paleti, co-Founder at Darwinbox, a leading provider of cloud-based Human Capital Management. Paleti discussed how businesses can harness hiring trends in 2024 to enhance recruitment, engagement, and retention, aligning with employee expectations and improving operational efficiency.
The Manila Times (TMT): Real-World Use Cases: How does Darwinbox apply “generative AI-powered analytics” in recruitment and retention? Can you provide specific examples of its application in identifying top candidates or predicting employee attrition?
Jayant Paleti: Here’s are some of the ways in which Darwinbox uses AI to streamline recruitment and enable organizations to identify and retain top talent:
- Org-Level Skill Gap Analysis: Darwinbox leverages a sophisticated native skills graph that maps out skills within the organization. This enables HR teams to conduct detailed skill gap analyses across different teams. Identifying these gaps helps in crafting targeted recruitment programs or development plans to bridge them, ensuring the organization is well-equipped to meet current and future job demands.
- Internal Talent Marketplace: Darwinbox’s facilitates a talent marketplace for employees by using AI to match employees’ skills and career aspirations with open positions or projects within the organization. This helps fill vacancies more efficiently with top internal talent and supports career development and retention by offering employees opportunities that align with their skills and growth ambitions.
- Identification of Critical Employee Segments: By using clustering techniques, Darwinbox’s AI can group employees into segments based on similar attributes, behaviors, or risk profiles. This segmentation allows companies to identify critical employee groups that may be at risk of low engagement or high attrition. Tailored strategies can then be developed to enhance engagement, satisfaction, and retention within these key segments.
- Predictive Attrition Risk Indicators: Through the analysis of various data points such as employee engagement scores, performance metrics, and other relevant HR data, Darwinbox’s generative AI models can identify patterns and signals that indicate a higher risk of employee attrition. By recognizing these risk indicators early, organizations can take proactive measures to address employee concerns, improve job satisfaction, and ultimately reduce turnover rates.
TMT: Addressing Bias: How does Darwinbox ensure that its generative AI models maintain fairness and avoid bias in HR assessments and recommendations?
Paleti: Here are some steps Darwinbox takes to maintain fairness and avoid bias:
- Diverse Training Data: Using diverse and representative datasets is vital to developing bias-free AI. In HR, this means training AI on data reflecting varied demographics to ensure fairness in hiring or employee assessments. By training on a strategically sampled dataset that mirrors the diverse workforce of a multinational corporation, our AI at Darwinbox can accurately predict employee turnover, considering the myriad factors that influence such decisions across different regions and job roles.
- Continuous Monitoring: Bias in AI is not always apparent at the outset and can evolve. 3.Regularly reviewing AI systems helps maintain fairness over time, a practice critical in HR tech. To ensure real-world variability is captured, Darwinbox makes sure any biases in outputs are addressed. For instance, when a Job Description (JD) is generated on the product platform using our generative model, the prompt to the model is engineered to avoid a basic set of biases by default, such as gender, race, and ethnicity.
- Eliminating Discrimination: In HR, it’s essential to configure AI algorithms to avoid discrimination. This can involve careful analysis and adjustment of AI algorithms to prevent discrimination based on gender, race, age, or other factors.
- Considering Edge or boundary cases: Training AI systems to consider edge or boundary cases ensures they are equipped to deal with unusual or extreme scenarios. For instance, in deploying AI for performance management systems in Darwinbox, considering edge cases allows identification and appropriate handling of outliers, such as identifying employees with atypical performance patterns due to unique circumstances, ensuring they are fairly assessed and supported.
- Regulatory Compliance & Ethics: In an era where data breaches are increasingly common, Darwinbox prioritizes protecting personal and sensitive information. Anonymization of data within our training pipelines ensures that individual privacy is safeguarded, a practice that reinforces our commitment to ethical AI development. This approach fosters trust between Darwinbox and our clients and ensures compliance with stringent data protection laws such as GDPR and CCPA. For our clients, this means peace of mind knowing their data is secure and their reputation intact. We do not make any calls to public LLMs such as ChatGPT and prefer using secure native LLMs that are custom trained with authentic anonymized datasets.
TMT:Beyond the Buzzwords: Could you explain the specific technological components within Darwinbox’s platform that enable “generative AI”? Which AI models or techniques are leveraged?
Paleti:Darwinbox has built one of the Industry’s only HR based Large Language Models (LLM) named PROSE (i.e People Relational and Organizational Semantic Engine) which is trained on various types of anonymized HR data such as Resumes, JDs, Policy Documents, Service Data, Feedback, Goal Descriptions, etc.
This powers several features such as:
The creation of content for job descriptions, competency guides, learning outlines, and onboarding and transition tools,
The creation of skills models, experience models, and candidate profiles for recruiting and talent discovery,
HR self-service and knowledge management,
Analyze and improve pay, salary benchmarks and rewards, and more.
Darwinbox has also developed its own Skills Ontology using PROSE that represents all relevant skills in a graph with dynamically changing distances between skills basis, organizational and industry context.
This centralized organizational skills framework provides a transparent, granular view of skills. Capturing a view of skills at an employee level enables an organization to unlock the full view of the employees’ capabilities, not just restrict the view to the current job the employee performs. The skills framework also enables organizations to fluidly shift skills across functions, departments, and other business units, depending on the business requirements.
TMT: On “boosting decision-making.” What kinds of HR decisions are particularly improved by accessible data visualization? Are there common problems resolved, or new types of actionable insights possible?
Paleti: Analytics in today’s day and age is all about uncovering the right insights at the right time and enabling more accurate decision making at every level of the organization. To that extent, Darwinbox, with the influence of generative AI, enables organizational leaders to ‘converse’ with their analytics dashboard. Questions like ‘why is my attrition rising in the Southeast Asia region’ are reverted with contextual drill downs, visualizations, and guided journeys to help enable better root cause analysis.
Darwinbox offers accessible reports and visualization that significantly improves decision making across the employee lifecycle with personalized dashboards across personas for talent acquisition, performance management, headcount and turnover, diversity and inclusion, employee engagement and more.
Furthermore, with employee adoption as a priority we have incorporated AI across reports and dashboards to democratize analytics and make data more accessible. AI backed smart nudges across these dashboards further help drive prompt action where necessary. And AI-powered smart summaries convert visualizations to more actionable insights.
TMT: Data-Driven Culture” Challenges: What are the biggest hurdles in shifting managerial mindsets to become more reliant on data, and how does your approach overcome them?
Paleti: Shifting managerial mindsets to embrace a data-driven culture is often met with several challenges. One of the primary hurdles is resistance to change, as managers may be accustomed to making decisions based on intuition or past experiences rather than relying on data. Additionally, there may be a lack of data literacy, where managers are not familiar with interpreting data or using analytical tools. Another significant challenge is the availability and quality of data. Managers may be hesitant to trust data-driven decisions if the data is perceived as incomplete, outdated, or inaccurate.
Darwinbox addresses these challenges by providing intuitive and user-friendly data analytics that make it easy for managers to understand and interpret data. AI powered smart summaries help employees quickly understand key takeaways from visualizations. Our platform offers real-time, actionable insights, ensuring that managers have access to up-to-date and relevant information. AI powered smart nudges alert employees of critical trends or events ensuring prompt action. Darwinbox also offers AI powered cross-module smart search where employees can type in their query, to quickly get insights at their fingertips.
TMT: Mobile’s Drawbacks: While mobile access offers convenience, what are the potential downsides of HR analytics on a small screen compared to a desktop? Are there trade-offs to be aware of?
Paleti: While Darwinbox as a platform is built to be device agnostic across mobile or web, here are some potential tradeoffs of HR analytics on mobile:
Accessibility vs. In-depth Analysis: Mobile apps offer real-time access to key HR metrics, allowing for quick decisions on the go. However, for in-depth analysis and exploring complex data sets, a desktop environment remains preferable.
Convenience vs. Accuracy: The ease of data entry on mobile can be appealing. However, organizations need to ensure data accuracy by implementing robust validation processes and potentially limiting the type of data entered through mobile apps.
Security vs. Transparency: Striking the right balance between security and transparency is crucial. While robust security measures are essential, it’s equally important for employees to understand how their data is being used and protected.
TMT: Talent Challenge Context: Why is high turnover particularly important for companies to solve right now? Are specific economic or labor market factors at play?
Paleti: We are experiencing the most entropic era in talent stability, ever. There are a lot of macroeconomic factors such as rapidly evolving markets, scarcity of skills, rising expectations of employees from their organizations, increasingly horizontal and diagonal growth expectations, ease of switching roles, and more. Employees are actively:
Reassessing Priorities: The pandemic caused many to re-evaluate their work-life balance and career paths. They’re seeking jobs that offer flexibility, purpose, and better compensation.
Facing Burnout: Intensified workloads and remote work challenges have led to employee burnout. People are seeking new opportunities with a healthier work environment.
Skills Gap: The rapid pace of technological change creates a skills gap. Employees with in-demand skills have more leverage and are more likely to be poached by competitors
The consequences of high turnover are far-reaching:
Loss of Productivity: When experienced employees leave, it takes time and resources to onboard replacements. This disrupts workflows and impacts overall productivity.
Increased Costs: The cost of recruitment, onboarding, and lost productivity due to high turnover can be significant, straining company finances.
Decreased Morale: High turnover can create a sense of instability and low morale among remaining employees, further fueling the cycle.
Competitive Advantage: Many HR platforms offer some form of analytics. What makes Darwinbox’s implementation of these concepts unique or more powerful compared to competitors in the field?
Darwinbox’s analytics are powered by cutting-edge AI and machine learning and a high fidelity and extremely scalable data lake, enabling predictive analytics and advanced data modeling. This allows organizations to not only understand current trends but also anticipate future challenges and opportunities. Furthermore, our platform integrates advanced analytics with a deep understanding of HR processes, providing actionable insights that are both relevant and strategic. We offer a wide range of analytics dashboards and reports, across the employee lifecycle, all within a single, user-friendly interface. This holistic view allows HR professionals to make data-driven decisions across the entire employee lifecycle, ensuring a seamless and effective HR strategy.
TMT: Implementation: For a company that’s not currently using advanced analytics in HR, how difficult is it to implement a platform like Darwinbox? Are there significant changes to HR processes that need to be accounted for?
Paleti: Implementing Darwinbox in a company that’s not currently using advanced analytics in HR is a straightforward process, designed to minimize disruption and maximize ease of adoption. Darwinbox’s analytics is built on the base of a robust data lake, that unifies all HR transactional data across the employee lifecycle and across multiple touchpoints including data from different Business Units and different modules. It also allows organizations to plug data from ancillary platforms such as financial systems, ERP, CRM, etc. via seamless integration on our native iPaaS layer.
Our platform is user-friendly and intuitive, ensuring a smooth transition for HR teams and employees alike. Darwinbox is designed to complement and streamline existing HR workflows, automating manual tasks and providing value-driven insights to inform decision-making. As a result, HR teams can expect to see improvements in efficiency and effectiveness, rather than having to adapt to significant changes. Our goal is to make the implementation process as smooth as possible, enabling organizations to quickly start leveraging the power of advanced HR analytics with minimal disruption.
Visit my column at Sunday Business & IT, The Manila times.