As a general rule, businesses should organize work into well-defined processes when the goal is creating large numbers of identical or nearly identical outcomes, as when Volkswagen needed to install the software that conned the EPA into hundreds of thousands of identical diesel Passats, Beetles, and Jettas.

Designing the software? That’s more of a practice.

As it turns out, there’s a third way to organize work that can, in some situations, give you the best of both worlds — the scalability of business processes without losing too much of the flexibility you get from business practices. We’ll get there in a minute.

But first, closure. The heating coil showed up Tuesday. Tracking number? Never got one. The service tech showed up Thursday and installed it … but didn’t need it, as he already had the replacement part in his truck. He was also kind enough to reverse the charges for the heating coil that was damaged in transit and delivered to North Carolina.

Thanks to all who offered empathy, sympathy, and advice for last week’s account of our adventures in Customer Elimination Management (CEM) and the Six Stupid Methodology so often used to implement it.

Underneath the six stupids enumerated last week is a single root cause: An intense desire to dumb down work until any gerbil can handle it. It’s a fundamental underlying assumption of well-designed business processes: Execute the steps in the proper sequence and quality automatically happens.

Cooks follow processes — recipes — to make meals. Chefs, in contrast, are practitioners of the art of haute cuisine.

Trying to turn customer service into a fixed set of steps and instructions is a recipe for customer elimination, as a certain unnamed retailer demonstrated to yours truly in the Case of the Burnt Out Dryer.

And “case” is the operative word: When a process has failed and customer service is needed, either (1) the customer service representative resolves the problem during the initial contact; or (2) the situation enters … or at least should enter … the realm of case management.

Case management is an example of a hub-and-spoke practice — a practice in which one person owns the situation and calls on whatever business processes, relationships, or resources are needed to resolve it. Had the first person we contacted in the course of attempting to repair our dryer assigned a case manager, it would have been a far smoother and satisfactory experience — admittedly a low bar to hurdle, but still.

Enter “Next Best Action,” which should probably be called “best next action,” but that ship has already sailed.

Very briefly, it’s a way to combine decision trees, rules-based AI, internal customer knowledge, additional customer knowledge gleaned from the social web, process tracking, and predictive analytics. The result is a system that replaces a fixed sequence of one-size-fits-just-a-few process steps with a flexible collection of possible actions, driven by a system that figures out what most logically should happen next.

In the case of our dryer repair, a next-best-action system would have learned that UPS redirected the initial part shipment from its initial destination in Minnesota to North Carolina due to in-transit damage, immediately ordered a replacement part, and notified us of the situation … my wife by text and me by email because it knows our preferred contact channels.

Having ordered the part it would have monitored the case for status changes, notifying us again when UPS picked up the part and assigned a tracking number.

And, it would have set a timer. When it expired and no tracking number was forthcoming, it would have assigned a human case manager to take it from there. Next best action doesn’t eliminate the need for human intervention. But it can narrow it down quite a bit.

Which leads to this potentially uncomfortable question:

Next best action is real. You can implement it before your competitors and take customers away from them, or you can implement it later and lose customers to them.

The question: Is your IT organization … your developers and business analysts … ready to implement it?

Let’s go one level deeper: Even if next best action isn’t in your future, something else new, different, and more sophisticated probably is.

Is your IT organization equipped to recognize it, incubate it, and put it into production?

We’re in the era of pervasive technology and IT has only three choices:

Lead, follow, or wonder what happened.

* * *

My expertise in next best action is distinctly limited. For a more in-depth account, look here.

Or, wait for the soon-to-be published The Cognitive Enterprise, in which next best action has a prominent part to play. My co-author, Scott Lee, has been implementing next best action since being part of GM’s OnStar design team more years ago than he’s willing to admit.