FIVE
The New Organization
Digital, Agile, Boundaryless, and Work-Centric
Have you ever played Jenga? Players take turns removing one block at a time from a tower constructed of fifty-four blocks. You then place each block you’ve removed on top of the tower. This creates a progressively taller and more unstable structure. Remember that feeling of removing one of the blocks holding a platform in the multistory tower and seeing the tower start to shake? Will it tip over or will it hold? Organization structures can be like the Jenga tower. You might be lucky, and changing one element, such as reinventing a job through automation, won’t affect the overall stability of your organization. More often, reinventing a job has collateral effects on other jobs, the relationships among them, and the communication, authority, and power structures of your organization. Often, if you’re not careful, reinventing only one job may make the organization unstable.
How can you identify or predict how job- and work-level changes affect the larger organization? Which reinvented job changes might even improve how your organization functions, and which are like the flaps of the butterfly’s wings that cause a tornado?
This chapter examines the organizational implications of optimizing work automation by reinventing jobs. Just as the blocks in the Jenga tower are decoupled from each other, the work elements and tasks that were once contained in a stable organizational structure of jobs will be increasingly decoupled from those former jobs and even from the organization. As we have seen, this unmooring of work tasks offers immense new opportunities to reinvent jobs in new ways. What are often overlooked are the new dilemmas and risks created when you remove the traditional safety net of job descriptions and organization designs based on them. It’s the difference between assembling a Jenga tower using a premade framework where you can only assemble blocks in a limited number of ways versus assembling the tower with unlimited freedom in the location and combination of the blocks, and when automation can change the very nature of every block at any time.
The four-step framework in chapters 1 through 4 helps you decide how to deconstruct and reinvent jobs. Those decisions also affect organization-level issues such as culture, diversity, alignment, engagement, authority, and accountability. Those organizational implications are often not obvious if you simply focus on reinventing single jobs. Sometimes, the organization-level implications will amplify the positive effects of work-level decisions. At other times, organization-level implications will resemble the tornado that collapses the Jenga tower. You may even decide to avoid reinventing some jobs, even though it seems logical at the job level. Savvy leaders must avoid being seduced by cost, risk, and productivity benefits from work automation applied to single jobs and consider the organizational implications.
The Outside-In versus Inside-Out Approach
There are two ways to think about the connection between work-level automation decisions and organizational implications.
First, you can work inward starting by understanding how automation might enable the transformation of your industry, strategy, business, and organization. Then, you consider how the work must change to support those organizational and strategic outcomes. A big idea from top officers often prompts this approach. For example, Tony Hsieh, the CEO of Zappos, seems to have had this in mind, as recounted in this excerpt from “The Unorganization,” a report by SogetiLabs:1
Tony Hsieh, the CEO of Zappos, sent an extensive email to all employees at the end of April 2015 explaining that they were going to be a company that organizes differently. The world changes quickly and business has become too unpredictable. Hsieh wanted to prepare his organization for this digital era. He wanted to put an end to old-school management, at least the last remains that were still present at Zappos. He said that everybody at Zappos must be adaptive, flexible, inventive and creative. In the email Hsieh announced that they would continue forward without managers. Self-organization was already a pillar of the culture at Zappos, but Hsieh’s ambitions went even further in implementing the so-called “Holacracy,” a self-organization methodology introduced in 2007 by the HolacracyOne company.
Executing such a vision requires aligning work and workers in new ways. Will you simply remove the work of the eliminated managers? Or, more likely, will some of that work be decoupled from the former job of “manager” and integrated into other jobs? Success will depend on optimizing work automation by replacing tasks previously done by managers, such as team communication, detecting exceptions that require attention, and analyzing and summarizing data from customers so that teams can quickly respond. The four-step framework for reinventing jobs in chapters 1 through 4 can be applied to issues like these to create the better-optimized new work necessary to support the organizational and strategic vision.
A second way you can connect the reinvention of jobs to your strategic challenges is to work outward by starting with job-level work-automation opportunities that enhance the cost, risk, and productivity value of existing jobs. Then, you must decide how new organization elements must support those newly reinvented jobs, or how negative organization-level implications might suggest reconsidering, slowing, or abandoning those job changes until the organization is ready.
For example, it is quite feasible to automate assembly-line jobs with advanced AI and social robots. However, at the organization level, production innovation ideas come from having humans on the assembly line so they can see ways to make the line more efficient, safe, or customer-friendly, and then communicate those ideas to operations engineers and designers. Remove the humans from the assembly line, and the line itself certainly becomes faster, more reliable, and safer. But, you lose the organization-level value of human assembly workers who find improvement opportunities and communicate those opportunities to production and product design teams.
Let’s see how each approach to organization-level optimization works. We will illustrate the outside-in approach with global appliance maker Haier. Then, we illustrate the inside-out approach by extending our earlier example of cancer surgery and treatment.
Working from the Outside In: The Story of Haier
Consider how automation changes the concept of a refrigerator.2 When sensors and AI can seamlessly and continuously connect it to the cloud, and those connections can, in turn, link to the cloud-based and sensor-powered systems of grocery stores, delivery vehicles, and suppliers (think Whole Foods as part of Amazon.com), the single appliance becomes a hub connected to a vast network of food suppliers. It becomes the entry point for “food as a service” that monitors household inventory levels, orders food, and delivers a fully stocked refrigerator and pantry. Such a transformation is only possible with a combination of technologies such as the IoT, cloud-based storage and services, and powerful cognitive automation and AI. Revenue now comes not only from selling the appliance, but from offering the best user experience in finding, acquiring, storing, and using food, not to mention the value creation and business opportunities arising from data collection, analysis, and interpretation.
How should we design an organization to address the challenges of such transformations? What happens when a manufacturing giant transforms itself to fully take advantage of these technologies?
Haier CEO Zhang Ruimin has become legendary for taking a traditional, hierarchical, global manufacturing organization and transforming it into a platform for serial entrepreneurship, with employees who act as self-governing entrepreneurs. According to an article in MIT’s Sloan Management Review, Haier has transformed from a manufacturing corporation to a platform providing financing, support, and coordination for a coalition of microenterprises, all focused on building products and services for the “smart home.”
Zhang’s concept is to turn customers into users who collaboratively improve and develop products by sharing their behaviors and ideas, and to reduce the distance between the organization and these users to zero, so that the organization is constantly co-creating with them. The enabling factors in this transformation are the IoT, a combination of technologies that allows information about the user experience to be constantly gathered, analyzed, and shared with the organization as it designs, develops, executes, and improves products and services that contribute to that user experience. Haier transformed its organization structure from a traditional hierarchy to a platform that encourages employees and partners to join microenterprises operating on the platform. The organizational platform has little hierarchy, but instead operates to provide support and market evaluation for more than two hundred entrepreneurial teams, each equipped with money, technology, logistics, and other support, and rewarded and empowered to act on the information they receive about the user experience in the smart home.
Compensation is determined by how much value you create for the users. As Zhang said, “When employees create value, they get paid. If they don’t create measurable value, they don’t get paid. Ultimately, if they don’t create value, they have to leave.” Instead of organizational units, divisions, product lines, and functions, Haier is organized as a goal-based entity, gathering and dispersing resources based on client needs. (
.)
As the CEO describes it:
The process begins with an objective. For instance, someone comes up with an idea for a product targeting a certain niche of the market. And then people from different departments or disciplines—research and development [R&D], sales, manufacturing, marketing—will sit down and analyze its viability across all the relevant dimensions. If they believe it is viable, they will form a community to bring it forward as a new microenterprise. Then they need to attach their plan to their compensation. We call it a predefined value adjustment mechanism, or VAM, which defines what goal the plan has to realize and how the members of the community will be paid if the goal is achieved. This is a signed agreement between Haier and its microenterprises. We also have microenterprises that focus on more cutting-edge projects. These teams may not plan to achieve revenue for a couple of years. Here, we set different targets and schedules. For example, at a certain point of this endeavor, they must be able to attract external venture capital. If they can’t achieve the investment by an agreed-upon time, then they have to let it [the project] go, or we might invite another entrepreneurial team to work on the project.
In our model we have delegated the major powers of corporate executives to the employees—or at least to the microenterprises—including the power of decision-making, the power of selecting and appointing personnel, and the power of financial allocation. Other companies would not do that. They believe that if these powers are delegated, managers will lose control. Our goal is different: We are trying to motivate employees to unleash their potential and realize their own value. We don’t want to control them.
The same “close to the user” approach underlies the role of business functions that provide services such as manufacturing. For example, Haier has 108 factories globally, each with multiple production lines. At Haier, each production line functions as its own microenterprise. Zhang describes it this way:
We evaluate the performance of these microenterprises based on cost, delivery and service quality, and market response to the products they make. This evaluation determines how they are qualified to get subsequent orders. Some production lines are able to acquire many orders. Some get fewer—and as a result, employees on those lines are not paid as well. Lines gaining more orders can merge with those having fewer. In this way the production lines are organically connected with the market.
Based on the aforementioned information Haier provided to us, we assessed the Haier case study, its relevance to this book, and the potential implications: the Haier example is a rather radical illustration of the power of technology to disrupt the concept of the organization. It is also an object lesson in how leaders must think about work and automation at several levels. At first, the issues may appear to be merely technological opportunities (AI, cloud storage, sensors, big data, etc.). As we have seen in earlier chapters, optimizing such opportunities requires deep attention to their impact and integration with the work. For example, the IoT allows AI and sensors to be built into refrigerators and other appliances, producing streams of data about operating conditions, but also about the consumer goods that are stored and used.
At a strategy level, the technological capability to link multiple products with Amazon-like services fundamentally shifts the very nature of strategy from making and selling great products to building the connected infrastructure that makes it effortless for users to manage their food acquisition, storage, and preparation.
Consider what happens at the work level in one job of a customer service representative. Such representatives now become AI-enabled. The old job included a work task of interviewing customer callers to discover problems. That work task is now eliminated with AI, augmenting the workers by allowing all representatives to start the call with much deeper knowledge about the customer’s problem. Automation also creates new work, because the appliance can actually track food usage and spoilage. Now the customer service representative can coach customers on how to use their personal device to connect the appliance to their Amazon or Alibaba account, so that they can place orders at Whole Foods or other grocery stores automatically. See how the work-automation framework from earlier chapters helps to work outside-in by starting with Zhang’s vision of “zero distance to the customer,” and reveals the implications at the work level that are keys to its execution.
You can also work outward from these work-level opportunities to see mid-level organization implications. The big vision is “zero distance to the customer” and an organization that is a “hub for data-connected entrepreneurs.” Between that vision and the work-design implications for jobs are vital relationships between teams, units, and functions. It is vital to connect the work-automation decisions described in previous chapters (that occur at the level of the work element or job) with the mid-level organizational consequences (that occur at the levels of teams, units, functions, and the organization). This is where issues such as trust, authority, accountability, information sharing, social networks, and cross-unit culture come into play.
For example, if we add the customer-coaching work described earlier to the call-center representative job, those representatives will discover ways to improve the product interface with Amazon and other cloud services. Is that the work of customer service representatives or product designers? Should customer service representatives now also formulate new product design ideas that arise as they coach customers on how to integrate the refrigerator with their Amazon account? How will Haier’s talented refrigerator designers react to the prospect that their work tasks will now be automated or transferred to customer service representatives? Maybe this work now becomes part of the refrigerator product design team and gives them the work task of coaching customers, augmented by AI. If Haier leaders rush too quickly to embed customer coaching in the call-center representatives’ work, they may inadvertently create conflict or territorial battles between customer service and product design teams, each of whom thought it was their job to tap the power of AI to work with customers to redesign products.
Taking this further, many Haier products, not just refrigerators, can now interface with Amazon. Perhaps the work elements that involve parsing and analyzing how customers use all the products in their home could migrate to a new functional division that never existed before. That division might now have the authority to oversee product design for all Haier products. It would contain work tasks such as listening to customers (or using AI to analyze customer interactions with service representatives) as well as working closely with Amazon or Alibaba engineers to devise creative ways to integrate all Haier products with those services. Again, what began as a work-automation analysis in the job of call-center customer service representatives now reveals questions about creating entirely new divisions, all because deconstructing the job liberated some work tasks to tap the power of AI.
These options require thinking beyond the work of customer service representatives, product designers, or Amazon liaisons. The decisions made at the work-automation level will have a ripple effect on organizational issues such as power, authority, accountability, information sharing, status, and culture. For example, a decision to create a new liaison division devoted to the Amazon or Alibaba relationship, spanning all Haier products, fundamentally changes the power and authority of the former product design group, which now will look to the liaison group for key information about product features and design opportunities. Product design leaders who formerly held the consumer-usage information may well chafe at what appears to be a structure that makes them subservient to this new division that is a liaison to Amazon and Alibaba.
Marketers and salespeople who formerly held great power and authority based on their knowledge about particular products and the channels that sold those products must now evolve to selling based on their expertise in understanding how multiple products connect with each other (such as the refrigerator, personal device, and microwave oven) and channels like Amazon and Alibaba. A leader considering work automation applied to sales and marketing jobs might use the four-step framework in earlier chapters to identify that the tasks of customizing solutions might benefit from work augmentation through better data from robotic sensors, information process automation, and analysis powered by AI. These are important, but not enough. Leaders must also consider the higher-level implications for the role of marketing work in the broader ecosystem of teams, units, divisions, and so on.
Working from the Inside Out: The Automation of Cancer Treatment
Recall our example of oncology treatment in chapter 3, and the compelling imagery of AI-enhanced reinvention of treatment identification and choice, and social-robotic surgical assistants enabling human surgeons to do less-invasive and more precise surgical procedures. In that chapter, we focused on reinventing jobs to identify optimal human-automation combinations. Table 5-1 shows the job-level reinvention of oncology treatment, as described in earlier chapters.
TABLE 5-1
Organizational implications of reinvented oncology treatment jobs
Job-level reinvention |
|
Old job |
Automation-optimized job |
Oncologist gathers and reviews patient information and estimates the likelihood of cancer. |
RPA gathers and integrates patient data in real time. AI reads the data, assesses the likelihood of cancer risk, and provides the oncologist with a preliminary score. |
Oncologist orders diagnostic tests, analyzes the results against knowledge of medical research findings, and makes a cancer diagnosis. |
Cognitive automation and AI (such as IBM’s Watson for Oncology [WFO]) scans millions of pages of medical literature and makes a diagnosis based on constantly upgraded algorithms. |
Oncologist evaluates alternative treatments and decides what treatment the patient will receive. |
WFO analyzes data more quickly and completely and makes recommendations on the more typical cases. Oncologist makes recommendations on cases that are unfamiliar to WFO. |
Oncologist carries out surgical procedures. |
Routine surgical procedures carried out by AI and sensor-equipped machines. Oncologist decides what tasks to delegate to the machine. Oncologist carries out the nonroutine tasks. |
Oncologist leads the diagnostic and surgical team, serving as a hub for information and decision making. |
Team can access data from multiple systems, supported by AI and RPA. AI alerts the team to exceptions requiring attention and tracks the behaviors of other team members. |
The Star Model of Organization Design
There are many frameworks, each with its own advantages and disadvantages. To structure our analysis of the organizational implications, we use the star model of organization design. This framework can help describe the organizational implications of work-automation optimization. The “star” model was formulated by Jay Galbraith and refined by thinkers such as Amy Kates, Greg Kessler, Susan Mohrman, Christopher Worley, Edward Lawler, and Stu Winby. Galbraith described the model this way:3
- Strategy: The strategy specifically delineates the products or services to be provided, the markets to be served, and the value to be offered to the customer. It also specifies sources of competitive advantage, or capabilities.
- Structure: The structure of the organization determines the placement of power and authority in the organization.
- Processes: Information and decision processes cut across the organization’s structure; if structure is thought of as the anatomy of the organization, processes are its physiology or functioning.
- Rewards: The purpose of the reward system is to align the goals of the employee with the goals of the organization. It provides motivation and incentive for the completion of the strategic direction.
- People practices: Human resource policies—in the appropriate combinations—produce the talent required by the strategy and structure of the organization, generating the skills and mindsets necessary to implement the chosen direction.
Galbraith could hardly have imagined the recent advances in work and automation. The star model was developed and typically applied to traditional organizations and work done in traditional jobs through employment relationships. However, these organizational elements can also apply to the emerging world of work and organizations. As we described in our book Lead the Work, deconstructed and decoupled work changes the very definition of fundamental ideas such as capabilities, structure, processes, metrics, and HR practices. The framework in this book suggests that “structure” and “process” are made up of increasingly deconstructed and decoupled tasks that are constantly reinvented and optimized to account for automation, as well as alternative work arrangements such as gigs, contracts, projects, tours of duty, and so on. It means that jobs and organizations are constantly evolving and reinvented. Table 5-2 describes examples of the necessary reinvention of the organization, using the “Star Model” elements.
Organizational implications of reinvented oncology treatment jobs
Surgeons Are No Longer Godlike
An oncology surgeon meets with hospital administrators to request more discretion for the surgical team, as they use diagnostic information and recommended surgical procedures produced by the robotic surgical assistant. The team likes the reinvented job that gives them deep patient information at their fingertips and using the robot for mundane things like opening and closing incisions. However, they bristle because now their reinvented jobs require them to conduct treatments and surgical procedures that follow the best practices identified by cognitive automation. Sometimes the surgical team believes different procedures would work better. Sometimes the cognitive automation recommends that the robot do a procedure because it has been proven superior, but the team members believe they are more adept. (See table 5-2 for the implications of reinvented oncology treatment jobs.)
Before automation reinvented these jobs, only the surgical team had access to information within the operating theater and personal expertise about patient reactions to surgical techniques. Before automation, surgeons were hired and rewarded for their personal expertise and unique capabilities to choose and perform complex surgery. Administrators relied on the perceptions, expertise, and information that surgical teams shared. Before automation, the only option for administrators was to give the surgical team maximum freedom to operate and make decisions, perhaps bounded only by broad guidelines based on cost or legal liability.
After automation reinvents these jobs, the administrators have access to comprehensive information about surgical practices, patient reactions, and the best practices of thousands of other surgical teams using the same AI and social-robotic equipment. The surgical team is now just one of many pivotal groups. In many cases, information analysts and technology designers and programmers have equally valid and important perspectives. As the surgical team evolves to become pilots of the AI and robotics, it is no longer optimal that they have the exclusive authority to recommend surgical treatment. The reinvented jobs of data analyst, programmer, or technology expert now have an equal voice in such decisions. That was unheard of in the past.
Automation has not just reinvented all of these related jobs; it has created an important challenge for the surgical center’s leaders and organization designers. They must balance these organization factors against the benefits and costs of work-level automation. Perhaps leaders should forgo some of the opportunities to reinvent jobs with automation that we described in chapter 4 and in tables 5-1 and 5-2, because the organization-level drawbacks are too serious. Perhaps the pace of automation must be slowed to allow surgical teams time to understand and acclimate to these new roles. Perhaps the optimum moment to implement automation is after the hospital has recruited and hired programmers and AI experts who also have surgical experience and qualifications, because they can bring necessary credibility and trust to discussions with surgical teams that traditionally held all the power and authority.
The New Organization
Automation reinvents organizations, just as it reinvents jobs. Such reinvention includes virtual teams, agile/SCRUM, holacracy, flat and lean organizations with self-managing workers, and the Haier example of the organization as a hub for microenterprises. Automation makes it possible and perhaps necessary to reinvent the organization and supports faster and more radical experiments. Smart software supports flexible staffing and human resource planning. Data analytics monitor performance and predict resource requirements. Smart communication software enables virtual collaboration. Augmented reality simulates in-person interactions. The IoT conveys vital customer and user data directly to the workers, with no supervisor.
But not only organizations are redefined through automation. The meaning of leadership and the role of leaders and followers are also reinvented. That’s the focus of chapter 6.