ACM Logo  An ACM Publication  |  CONTRIBUTE  |  FOLLOW    

What is workflow learning?

By Jay Cross / July 2004

Print Email
Comments Instapaper

A buff venture capitalist in a designer suit steps into my elevator. Soon she asks, "Workflow learning? What's that?"

I reply: "That's something you won't have to ask five years from now, for by then Web Services and the integrated, real-time enterprise will be commonplace. Learning will have become a core business process. It's what will connect humans to their work.

Now, we call it workflow learning because it's a different animal from the workshops, reference manuals, computer-based training, and courses that have been the staple of corporate training departments. The goal of workflow learning is to optimize business performance. It employs "smart" software to guide, inform, and assist workers to do their jobs better. When appropriate, it will put the worker in touch with the right expert or mentor or help desk, someone who's both knowledgeable and available.

You've heard about the promise of Service-Oriented Architecture? XML? It's finally happening. Intelligent software agents tackle the Mickey Mouse work, freeing workers to serve customers, make judgment calls, generate innovations, and accelerate the pace of the business. IBM calls it "on demand;" at HP it's the Adaptive Enterprise. This will spark a sea change in the way the world conducts business.

Finally, we're shrugging off the practices we inherited from the industrial era and entering the knowledge economy. Workflow learning is the people part of that."

Describing workflow learning invokes the corollary to Murphy's Law: "You can never do just one thing." In this case, you can't appreciate workflow learning without understanding its context. Critics and reactionaries write off the workflow vision as futuristic and not yet relevant. To me, it's more what William Gibson meant when he said, "The future has already arrived. It's just not evenly distributed yet." Companies are modeling workflow now. Workflow engines are at the heart of IBM, Microsoft, Oracle, BEA, SAP, and PeopleSoft.

Competition is driving business cycles to occur faster and faster until "soon" is replaced by "right now." In a zero-latency enterprise, planning collapses into action. Successful organizations identify and demolish obstacles in the path of slack-free workflow. Frills and inefficiencies such as studying for the test instead of the job, going through the motions thoughtlessly, and fighting the last war largely disappear.

Rigidity, discipline, narrow focus, and top-down control improve efficiency when change is rare. This logical formula worked so well in producing physical goods that it became gospel, seared into our thinking as the one best way to do things no matter what. But the world that created the Industrial Age is long gone. The vestigial worldview of the old era is wholly inappropriate for our current situation.

Today the goods are largely intangible; innovation trumps obedience; flexibility is the only way to survive; and control is being delegated to the workers themselves. In fact, in a business ecology, the workers are in charge, and most organizations are bossless.

Technology, communication, information, world population, and interconnectedness are advancing exponentially. In the face of such volatility, long-range planners are on permanent leave of absence. In some quarters, "long-range" means "next month."

Christopher Meyer and Stan Davis (2003) support the notion of workers being in control in their book It's Alive: The Coming Convergence of Information, Biology, and Business. They point out that "The past fifty years have shown conclusively that distributed decision-making does a better job of satisfying demand than a centralized approach. Nonetheless, many of our businesses retain a surprisingly `Soviet' management style, using approaches developed in an assembly-line era that have more in common with a top-down mentality than with a bottom-up one."

In a workflow learning ecology, each worker has an instant picture of his or her slice of the workflow on the screen 24/7. Assuming the right permissions, the worker can also realign the process, correct error situations, receive a chunk of simulation-based learning, or be connected with another person. The "dashboard" has become the "cockpit." As network computing becomes pervasive, the cockpit may reside on a handheld, a head mounted display, or a wi-fi bodysuit.

Skeptical? You're witnessing bottom-up decision-making in the swift turn of a school of fish but we're not used to seeing it in business. MIT Professor Tom Malone equipped an audience with hand-held paddles whose position could be read by a computer. On the screen up front, he projected a flight simulator. A hundred people jointly took the controls of the plane and, against all expectation, flew the plane without a pilot and without a crash

When I attended business school—admittedly that was soon after the Dark Ages—I took separate courses in marketing, accounting, finance, organizational behavior, international business, and so forth. Then I joined a corporation that had a marketing department, an accounting department, a finance department, etc. These separate entities competed for senior management attention when they should have been optimizing corporate performance. Had this outfit lived in nature, it would not have survived its first day. A business ecosystem must adapt to fit its circumstances; the elements of a business ecology communicate through Web Standards. In addition, a business ecology doesn't have a boss; it grows through the collective effort of its workers.

In the past, if a business wanted to automate, it had to pretend its organization operated like a machine. Now however—for the first time—IT is sufficiently flexible to conform to the shape of the business it serves. Under the hood of the future computer you'll find a bevy of snippets of agent software working out solutions—without the heavy hand of a programmer to hardwire them together. Since the software agents are "loosely coupled," they enjoy the flexibility to reconfigure themselves for best fit.

Another integral factor in business is intangibles. Forty-percent of the workers in America are knowledge workers, and their numbers are growing. They don't make anything that classic accounting considers an asset, yet corporations keep paying them high salaries. One definition of intangible is "assets that are saleable but not material." A second definition is: "lacking substance or reality; incapable of being touched or seen." Finally, another definition is "not having physical substance or intrinsic productive value." Whoa. That sounds close to worthless. More than 25 percent of the net worth of American public companies is intangible. Don't tell me that it's worthless.

The Coca-Cola Company has a market value of $164 billion. Its tangible assets are worth $7 billion. The other $157 billion is intangible. It's the brand, the social capital, and know-how that Coke has built up over the years that are really valuable. Similarly, all but $20 billion of IBM's value is intangible; reputation is intangible. Microsoft has only $1 in tangible assets for every $3.75 reflected in its capitalization; code is intangible. This is not just some financial anomaly. Clearly, intangibles have a tangible impact.

The discipline of Business Process Management (BPM) provided a dominant gene in the DNA of workflow learning. BPM is all about manipulating intangibles. It is inevitably a cohort of workflow learning. Paul Harmon, author of Business Process Change: A Manager's Guide to Improving, Redesigning, and Automating Processes, defines BPM as "aligning processes with the organization's strategic goals, designing and implementing process architectures, establishing process measurement systems that align with organizational goals, and educating and organizing managers so that they will manage processes effectively." In other words, you draw a chart of the flow of work though the organization. Now for the magic: Resequence or redraw something on the diagram, and the system rewrites the underlying code. Thus, you manipulate the symbols and you change the process on the factory floor. At the same time, you can imbed information for the worker who later gets stuck at this point. Or you can ask the software to generate a simulation of the ways things should operate.

For some, the work of the future will resemble an elaborate, personalized video game front-end that's connected to the physical operations of their company.

Life will be simpler five years from now. We'll be comfortable living in a world of organizations without bosses, computing without programmers, and webs without weavers.

Next time, if a corporate titan comes into my elevator and asks me about workflow learning, I'll simply tell her that workflow learning is how workers improve performance in a business ecosystem.


  • There are no comments at this time.