The value of HR analytics stems from a strategic decision

Without understanding and applying the paradigms of Artificial Intelligence, HR Analytics would simply be dashboards for data visualization — useful for providing real-time summaries of the many reports that HR departments have always produced.
For many organizations, this is certainly a step forward. But the value created is limited to improved accessibility and consultation of performance results, talent reviews, related nine-box grids, and succession planning tables.

By contrast, the approach of creating Artificial Intelligence solutions for HR allows us to extract meaningful insights from available data and information to support HR action.
In particular, AI solutions for HR enable us to:

• Describe: for example, explain the revenue of a retail store based on a range of factors that human intelligence alone cannot analyze.
• Diagnose: for example, explain inverse correlations between the number of FTEs and store revenue — i.e., answer the question: "How is it possible that revenue increases even with fewer staff than standard?"
• Predict: continuing the example, identify that revenue could increase at a specific location if certain conditions relating to people, teams, or location are met — regardless of the number of FTEs.
• Prescribe: determine where to open the next store, what kind of staff to hire, and how to build the team.

AI solutions for HR serve as a form of Business Intelligence support to proactively identify areas for action and enable fact-based decision-making.

These results can be achieved by adopting an applied research approach to Data Science:

1. Define what matters: this phase may require significant investment. In our view, it involves both HR Experts and Senior Leaders defining the key questions that data should help answer — questions that integrate both People and Business perspectives.
2. Data preparation: preparing the dataset with data categories aligned with the defined questions. It is critical to ensure data availability, accessibility, quality, and reliability. For more on this, read here.
3. Model training: in this phase, a training dataset is analyzed using algorithms under supervised or unsupervised learning logic. The goal is to identify the most appropriate algorithms to answer the questions from phase 1. A validation phase using a test dataset follows.
4. Gather insights: at this stage, results are analyzed across the entire database using the algorithms refined during training.
5. Run: the solution is implemented and can generate dashboards with HR Analytics, potentially automating certain decisions.

It’s easy to see that a single HR Analytics software cannot enable HR departments to create indicators aligned with their unique objectives and organizational characteristics.

Instead, it is necessary to bring together the business perspective with data intelligence to identify the right challenges and leverage data’s potential to address them.

If you think about it, the HR and IT departments are the only ones capable of "seeing" and understanding the entire organizational system. It’s no coincidence that these are the two departments most frequently called upon — and often volunteer — to lead Digital Transformation initiatives.

The importance of a systemic approach is, therefore, self-evident.
But what does it mean? And how can it support the development of Data Intelligence solutions for HR Analytics?
HR Analytics must offer a representative view of organizational phenomena in order to support decision-making on Human Capital matters. We believe Human Capital decisions must not be separated from business decisions.

This strong belief is rooted in the emerging trends in the evolution of work, which show a growing shift towards decentralized, network-based organizational models.
Traditional performance review models are being replaced by real-time feedback mechanisms. Leadership roles are being redesigned to operate closer to the customer and further from the untouchable Headquarters.

This shift is driven by the need for speed and agility in order to stay competitive. Organizations need to increase trust and empowerment across all people in order to ensure timely, on-the-ground strategic decisions.

Currently, most available metrics are quantitative only — predefined KPIs built around a static organizational design. HR typically works with siloed reports from each HR process, which cannot measure the phenomena typical of new organizational models.
Traditional structures built on lines of business, with clearly defined roles and scopes of responsibility, cannot ensure future — or even current — competitiveness. These are being replaced by network-based teams and Agile models.

In this context, it’s no longer possible to rely on predefined measurement models.
Only through the ability to determine what matters, build heterogeneous data sets, and foster collaboration across all departments will it be possible to develop effective algorithms that provide HR with valuable insights for People Management and Business decisions.

This requires exactly the kind of applied research approach that, through Artificial Intelligence, can deliver real-time descriptions, diagnostics, predictions, and prescriptions of ongoing phenomena, and support HR decision-making.