This chapter covers two presentations discussing additional considerations for facilities staffing modeling. The first presentation highlighted the role that artificial intelligence (AI) could play in human resources modeling and the business implications of AI use. The second one considered the importance of incorporating change management practices into the implementation of workforce planning initiatives.
This workshop session focused on the increasing role of AI in the workplace, its effects on employees, and its potential implications on business outcomes. For this session, the committee had asked the speaker to address the following questions, which he touched on throughout the course of his presentation:
What kind of data would be of particular use when using an AI approach to staffing modeling? What are the applications and limitations of AI in forecasting human resources needs? What are examples of workforce planning problems that are best solved using an AI system? And how can AI build a system that evaluates thousands of possible forecasting models and chooses not only the method that is best, but which subset of the thousands can be best combined in an ensemble of models to increase forecasting accuracy?
Anshul Sheopuri (IBM) provided background on AI and discussed IBM’s journey as an organization that has attempted to implement AI within the company. He asserted that when AI is adopted within an enterprise, it results in improved business results, which are often reflected in better margins, better revenue, and better business metrics. However, he said, the biggest challenge of adopting AI lies in the organization’s ability to integrate AI solutions into existing processes and systems. In order to be effective, AI tools must fit into the existing workflow, incentives must exist to use the tools, and the tools must drive business solutions.
Artificial intelligence, explained Sheopuri, isn’t new—the term was first coined in 1956. After a long “AI winter,” there has been a resurgence of interest in AI due to such factors as the decreased cost of the technology, the explosion of available data, and the increased sophistication of AI algorithms. As a result of this resurgence, he said, in certain narrow domains AI solutions now are equal to humans in decision quality. He noted, however, that most experts predict that it will be 30 to 50 more years before general AI, across domains, will be at this same high level of decision quality.
Along with the recent resurgence of AI, Sheopuri explained, the marketplace is concurrently experiencing significant disruptions in standard business models. As a result of these global changes in the marketplace, the talent makeup of even such “old school” industries as oil and gas is drastically changing to include employees with diverse skill sets, such as design experts, product managers, and data scientists. Along with this change is a rapid decline in the half-life of skills, requiring more frequent retraining of employees. In addition, the marketplace changes have led employees to have higher expectations for their workplace experiences: “The way they experience movies or the way they buy goods at home, those experiences are now the new North Star. It is no longer okay to have clunky experiences within the workplace and exceptional experiences at home while you’re watching a movie or buying diapers on a webpage.”
These disruptions in the marketplace, noted Sheopuri, are occurring simultaneously with an explosion in the number of AI tools and technologies that can be used to manage various aspects of human resources, including recruiting, hiring, staffing, recognition, performance management, and leadership and development. Enterprises will have to determine how to navigate this new technological complexity in order to transform their practices and achieve positive business outcomes in this new environment.
As AI tools and technologies are implemented in an organization, Sheopuri emphasized, validated user research needs to be performed to understand their effects on employees, managers, and business leaders. He explained that IBM has implemented such user research over the past 5 years, which has helped the company to successfully adopt various AI tools, reducing failure rates or helping them to “fail fast” in a few months’ time instead of after 12 or 18 months of investment, or to scale up the use of the tools quickly if they are successful.
Sheopuri next turned to how IBM has successfully embedded AI into its human resources function, using these tools across an employee’s life cycle, from onboarding to development to operations. In this process, information from an employee’s resume and from training and performance is collected into an aggregate view that shows that employee’s level of expertise along specific skills. These skills data are available to everyone, including the employee, who can update the data. Sheopuri explained that these data provide IBM with a way of understanding the skills of its employees and the progression of those skills across the entire workforce. The data also allow employees to engage in their own improvement and help business leaders to make better decisions. This integration of AI into human resources processes, and the use of other AI technologies such as digital assistants, have improved the employee experience, which ultimately can lead to better business outcomes. Sheopuri stated: “The notion that you can really make an employee’s life better, make them more productive, help them be more effective in making decisions, and that could drive client experience, is something that we have learned over the past few years can be very productive and powerful in driving business outcomes as well.”
As IBM began to embed AI tools and technologies into its processes, Sheopuri said, the company discovered that its employees’ skill sets did not have the capacity and depth needed to help them most effectively leverage the new technologies. In response, IBM incorporated new, relatively inexpensive ways of upskilling, with self-serve content that is directly embedded in each employee’s workflow. Sheopuri noted that the upskilling effort has begun to show results and that employees are self-motivated and self-directed when they have these learning opportunities. These motivated employees tend to seek out collaborative, self-directed ways of working that provide them with regular feedback, and self-motivated, short-segmented methods of working have translated into reduced failure rates for some of IBM’s products.
Along with implementing AI tools in the human resources setting, Sheopuri said, IBM has also undergone process transformation, in which the organization has transitioned from thinking about processes like the recruiting and onboarding of employees in terms of silos and begun to focus on the overall employee experience. He noted that the recruiting and onboarding processes, and other processes that involve mundane, process-centered work, can be enabled with chatbots. Chatbots both improve the employee experience by allowing new employees to engage with the enterprise up front, and free up human resources employees to perform higher-value work. He noted that these process transformations have led to faster process speeds and better business outcomes.
Cheryl Paullin (committee member) asked Sheopuri whether AI approaches have been used to determine how many of certain types of employees are needed to effectively manage workload. Sheopuri replied that IBM uses technology to understand the extremely diverse skill sets of its 350,000 employees. He explained that in the past IBM determined staffing by identifying business objectives related to revenue, profit, and associated costs
and mapping those factors to the headcount needed to deliver the desired outcomes. According to Sheopuri, this method resulted in inaccurate projections because it was not taking skills into account. Today, he continued, IBM tracks 5,000 skills at a high level of granularity, using a skills management tool called Employee 360. This approach allows IBM to assess not only the skills, but the depth of each skill that is necessary to achieve desired business outcomes.
Paullin also asked Sheopuri to expand on the creation of chatbots. He responded that the technology has evolved so that bots are fairly easy to create, even by those who are not software engineers. Domain experts in the organization, who understand the types of questions the bot might be asked, can “train” the bots in 2 to 3 weeks. The technology can then figure out alternative ways that same question might be asked. Sheopuri encouraged the use of bots, stating that in his opinion bots “can drive massive productivity gains and employee experience gains in many settings.”
To conclude, Sheopuri shared what he sees as the main takeaway from IBM’s use of AI: the technology is only a building block, and it will not drive business outcomes in the absence of a more holistic transformation of an organization. He concluded: “What we have found is that, when you only focus on the tools and technology without these other building blocks, without upskilling the teams, without new ways of working, without clarity on the offering strategy, without having a data platform that is fit for purpose, what ends up happening is you have a good shiny object, but that doesn’t necessarily translate into business outcomes.”
This workshop session covered change management practices to drive successful workforce planning. For this session, the committee had asked the speaker to address the following questions, which he touched on throughout the course of his presentation:
What change management principles and techniques should be considered to ensure decision makers are willing to buy into the workforce planning premise and act on their findings? What are common roadblocks and strategies to overcome them? And, what are some lessons learned from the implementation of a workforce planning initiative?
Robert Motion (Raytheon Company) said that change management has been a common thread through the three positions he has held over the past 10 years of his employment with Raytheon Company. As a defense contractor with 65,000 employees, 40,000 of whom are engineers, Raytheon’s workforce planning primarily involves the needs of the company’s technical talent.
Motion noted that change management is a very large part of successful strategic workforce planning, and he posited four common change management challenges: creating an aligned understanding of what workforce planning will look like; demonstrating the business value of workforce planning; proving the impossible, in terms of providing a vision for what the future will look like with workforce planning; and building an army to manage change through an understanding of all the stakeholders involved in the change process.
In terms of the first challenge, Motion noted an organization’s leaders often do not share the same understanding of what workforce planning is and why it is needed. An aligned understanding has to be reached regarding the specific purpose of workforce planning in the organization, and he stressed that a clear definition of what workforce planning involves helps to align stakeholders. Motion explained workforce planning as a human capital story that involves “having the right people in the right place at the right time and at the right cost.” An organization has to understand and be in alignment on its human capital needs, in terms of capabilities, location, and timing, and those factors need to be rooted in analytics and data science, not simply based on gut feeling. The plan developed to achieve the desired workforce must be strategic and deliberate, and the effort must be collaborative and cross-functional, he said.
Motion then discussed four key considerations for meeting the second common change management challenge of demonstrating the business value of workforce planning: executive championship, a strong business case, benchmarking, and scoping. Executive championship is critical for the success of this process, he said: strong backing of a workforce plan drives commitment and follow-on. For a strong business case, metrics that tie the impact of
the workforce planning initiative to revenue and cost are essential. Benchmarking against peer organizations in the industry is another way to demonstrate business value, Motion said, because benchmarks help an organization to understand the standards and best practices to which it should be adhering. In terms of scoping, he noted that trying to do everything at once is not a good solution. Instead, he suggested using an incremental approach that will help long-term initiatives not run the risk of losing momentum. Raytheon prefers an incremental change approach to workforce planning, which provides “quick wins” that establish credibility.
By proving the impossible, the third change management challenge, Motion said he means telling people what the future will look like with workforce planning in place. Through benchmarking, a proof of concept can be created that will enable stakeholders to see what success in workforce planning could look like, in terms of strategy and business implications, the roles and competencies making up the workforce, and current costs compared with future costs.
At Raytheon, Motion’s team built a workforce planning toolkit to help with implementation, which included a workforce planning “playbook” to assist with training on workforce planning; interview protocols to enable proper workforce segmentation; decision analytics on the roles of interest, including data on such aspects of the workforce as voluntary turnover rate and retirement rate vulnerability; and forecasting templates to relate financials with roles, head counts, and staffing needs. Also, in terms of predicting the future, Motion stressed using the organization’s inherent strengths. In the case of Raytheon, where the strengths are analytic, this meant leveraging data to “talk the language of the leaders” and help the workforce understand and engage with the workforce planning initiative.
Turning to the final common change management challenge, building an army to manage change, he reminded the workshop participants that people fear change and recommended a book by John Kotter and Holger Rathgeber1 that uses the format of a fable to illustrate the fact that stakeholders will have different wants, needs, and attitudes in a change situation. He highlighted the importance of crafting both the workforce planning solution and the communication with stakeholders around the wants and needs of stakeholders.
Motion described the types of stakeholders—the army—needed in a workforce planning initiative, who he called bankers, customers, consumers, and teammates. “Bankers” are the people who fund the work, and Motion noted that the banker is often the executive champion of the project. The banker needs to understand how the change will add value to the organization. “Customers” are the business leaders who will “buy” the product, which, in the case of workforce planning, are the people who will spend time on the activity. Customers need to understand how the change will both mitigate risk and add value, based on hard data. “Consumers” are the human resources staff, the people who will use the product. At Raytheon, Motion noted, education of the consumers was very important, particularly in the areas of data, statistics, and models. Since human resources staff are the conduit to the leaders of the organization, Motion stressed that they need to be fully on board. “Teammates” are the individuals helping to implement workforce planning. Teammates, Motion said, need to be educated like consumers, but they must also understand the vision for the organization and help to drive alignment. Successful workforce planning has to integrate the various perspectives of stakeholders to create an aligned plan that will “build an army” in support of the proposed changes. In Motion’s words: “The idea here is, by putting this together as a collaborative cross-functional team, we end up with one voice across the organization.”
Motion listed four key capabilities needed in order for human resources to support workforce planning effectively: data acumen, general business acumen, partnering skills, and change management skills. Motion’s team performed a self-assessment of the human resources workforce in terms of these capabilities and then instituted immersion learning to upskill them on data analytics and business acumen, areas in which they showed less capability than other areas.
Fred Switzer (committee member) asked Motion how the labor market intelligence included in Raytheon’s workplace planning toolkit was obtained. Motion explained that some of the data were obtained through location-specific analysis of Raytheon’s primary business across the United States. As a long-term strategy, location-specific analysis of such factors as cost of living, availability and cost of skills, and competitive presence can be used to determine the best locations to expand a business. With shorter-term decisions, such as whether to “make” or
1 Kotter, J., and Rathgeber, H. (2016). Our Iceberg Is Melting: Changing and Succeeding Under Any Condition. New York: Penguin Random House.
“buy” employees, these data help the organization know where pockets of talent are located. Motion noted that such tools as LinkedIn, labor market analysis tools such as Burning Glass and TalentNeuron, and publicly available information are also used to create that part of the toolkit.
Paullin asked how Motion’s human resources team developed its expertise in change management. Motion explained that Raytheon has an organizational effectiveness department, that, at the time of the described workforce planning initiative, had recently developed a very robust change management curriculum specific to the company. So his human resources team “essentially attended a crash course in change management.” He noted that Raytheon also worked with an external consultant to get an outside perspective and that she was a very provocative voice in helping to drive the change. Stakeholder risk assessments were also implemented to determine who was on board with the initiative and who still needed information. Motion said, “It was a combined effort. It was, in retrospect, a really powerful combination.”
Alberto Galué (committee member) asked about validation of the workforce planning model. Motion outlined two different approaches undertaken over the years by his team. First, a revenue forecast was built, which was historically accurate, but it fell apart in terms of staffing needs and they had to learn to account for the shifts. They did so by introducing a Monte Carlo simulation, which provided a best-case estimation in terms of staffing needs at the skill-set level. This output was used for their recommended plan line. Motion added that historical staffing data were also studied, and a shift in the mix was incorporated into the overall model, based on the skills needed for the particular types of business projected.
In summarizing, Motion reiterated that change management was ultimately more important than the technical workforce planning solution. He stressed that workforce planning is a journey, that alignment across the organization is absolutely critical, and that starting small and generating initial wins are key to credibility. He again emphasized the importance of a continued focus on business value throughout the process, as well as of leveraging the strengths of the organization and building an army of advocates to support a workforce planning initiative.
This page intentionally left blank.