Analyst Perspectives

Roughly half of my more than 30-year career in human capital management was spent as a line manager responsible for HR technology strategy, selection and deployment. I learned a number of lessons during these years — some just in time, some after the fact. If I had to identify one common thread that unites these insights, it would be that inadequate attention to change management is an ROI-killer on these strategic initiatives every time.

I recall, for example, a particularly rocky HCM systems deployment early in my career. One of the many realizations my team drew from that experience was that unreliable data — or more precisely, the perception that data might be unreliable — undermined broad-based adoption, which in turn undercut the return on the investment. Causes for unreliable workforce data include not only subpar processes for data collection and validation but a lack of data standards, for example, rules for how terms such as status are defined and how appropriate values are assigned or derived.

After the initial roll-out to domestic business unit heads and their line managers, the team concluded that subpar adoption was chiefly caused not by system functionality issues but by distrust of the HR data. So before we deployed to other countries we focused on effective education as the foundation of an effective change management plan of attack. It involved virtual training and Q&A sessions on data definitions and data validation steps. Not only did data quality improve immensely, but confidence in the new platform improved. That in turn drove up adoption and therefore ROI.

More than one-third of the HR practitioners surveyed by Ventana Research in our benchmark research on Next-Generation Human Resources Management Systems identified improving accuracy of HR data as a top priority for HRMSs, yet many organizations don’t sufficiently emphasize the role people play in data quality. Education on data quality should be a major aspect of any HR-technology-related change management program.

Rapid advances in business technology, particularly in AI, mean there’s no shortage of opportunities for organizations to improve operations and overall business results. These opportunities include burgeoning use cases that rely on AI to predict employee-related outcomes and behaviors, prescribe actions (such as when an employee is a flight risk or a manager shows bias in employee-related decision-making), personalize aspects of the employee experience, and analyze workforce social or influencer networks, or even sentiments in aggregate.

However, without an effective change management program, organizations likely will struggle to navigate the path from system selection to successful deployment and implementation. Even the best-chosen technology won’t deliver on its potential without effective change management. To take full advantage of emerging digital capaibilities in HCM technology, an organization must plan and execute a change management program that includes three critical steps.

Assess organizational readiness.

Any serious attempt to assess organizational readiness for a major HCM systems change, which is one of the initial undertakings of change management, includes identifying and inventorying the new skills and competencies that have emerged in the digital era. Many organizations will conclude from their readiness assessment that they are short on experience in analytics and predictive models, and that they likely won’t have large contingents of data scientists any time soon. Organizations should nevertheless strive to make analytical thinking a core competency across the enterprise and assess readiness in that regard.

Identify potential sources of resistance.

One of the ways that HCM systems are getting more sophisticated is in their support for agile organizational structures. This is sometimes referred to as a matrix organizational structure with employees potentially having multiple reporting lines, but it’s more than that. It can also describe an environment where employees perform multiple roles across multiple organizational units or projects. This makes it more challenging to monitor support for or resistance to a new HCM system.  

AI-enabled sentiment analysis tools can help organizations identify problems by tracking prevailing attitudes such as receptivity to the planned change. A key consideration here, though, is that few variables that influence resistance to an initiative remain constant in agile organizational structures. Supervisors and team members significantly influence attitude and satisfaction at work and roles can be more fluid in these agile structures.  

Tailor the case for change.

Before adopting HR technology infused with AI or other forms of cognitive automation, organizations must evaluate how easily people will relinquish control to systems. This could mean that systems will do more data handling — scrubbing, validating, mapping and consolidating — based on business rules and pattern recognition. While some stakeholders may welcome this change, others will be inclined to trust their existing data quality assurance processes more, particularly those in payroll departments in light of the “zero-tolerance-for-errors” mantra of most such units. These tried-and-true processes may even include the longstanding “eyeballing” method, particularly if people have relied on this approach for years. This is where the notion of tailoring the case for change comes into play. Automation can improve accuracy but it won’t work when groups are wedded to their own data quality assurance methods. In these situations, for the change to move forward it’s important for the HR or HR technology team to establish agreements with operations areas such as payroll departments on when to delegate tasks to machines.

In the past, resistance to a new HR system often has been rooted simply in the desire (or determination) not to learn to use yet another piece of technology. Today, though, ominous prognostications about the automation of the workforce have given rise to new anxieties about digital innovations in HCM. These misgivings relate to fears about the extent to which machines will start making decisions about employees without human involvement. The fact is that this will not be happening anytime soon. And there are other concerns about systems monitoring email messages or SMS interactions, or about how quickly it takes the machine to reliably learn patterns and relationships before prescribing anything. These attitudes underscore the need for a much larger and longer commitment to helping system users distinguish between fact and fiction, as both are pervasive in the present HCM systems landscape.

Organizations looking to be an early adopter of AI in HCM and take advantage of the many potential business benefits these initiatives can deliver, whether personalizing the employee experience or predicting flight or compliance risks, should realize that effective change management is essential for success. This includes assessing organizational readiness and support levels, putting in place key staff, competencies and processes, identifying pockets of resistance and tailoring the case for change. The bump in adoption and ROI will justify the effort.