Double, double, toil and trouble: use AI to reduce toil, not value

Cortex newsletter by Jason Bloomberg

In the familiar incantation from Macbeth, the three witches seek to double the toil and trouble in the upcoming tragedy with their magic.

In marked contrast, organizations have long sought to reduce toil in their organizations – as well as all the trouble it causes.

The concept of toil (but not the term itself) entered the realm of management philosophy as early as the 1940s with the concept of Muda, or waste, in the Toyota Production System in Japan.

The system’s principles evolved into the broader Lean management philosophy in the 1990s with the books The Machine That Changed the World and Lean Thinking.

The term ‘toil’ entered the management lexicon in the 2000s, and by 2016, appeared in Google’s Site Reliability Engineering book, which applied the term to IT operations.

It didn’t take long for people to pick up on the concept, extending it to DevOps, process automation, and other transformational technology trends.

Today’s trend, of course, is AI. Reducing toil remains a priority – but so much of the pushback against AI results from its application well beyond this Lean principle.

We’re approaching an important crossroads. As the resistance to all things AI expands, organizations looking to leverage the technology to achieve lasting, strategic business value should look to the lessons Lean teach about toil to ensure their AI efforts aren’t in vain.

A primer on toil

Toil refers to work that consumes time, resources, and effort without creating meaningful value, in particular for customers.

Common examples of toil include:

  • Repetitive administrative tasks
  • Manual data entry
  • Duplicated work
  • Unnecessary process steps
  • Correcting errors and other rework
  • Various examples of waiting, including waiting for approvals or for additional information or resources.

Automation has long been the stock solution to the problem of toil in organizations. If a computer can handle a sequence of tasks on its own, it can take those tasks out of the hands of humans, thus reducing their toil.

AI, in particular agentic AI, has now become the go-to choice for automation, because its language model-based capabilities can automate a range of tasks that heretofore required human input.

Therein lies the problem: AI has become so capable that its scope can extend well beyond reducing toil to tasks that provide meaningful value to customers, employees, and others – taking those tasks out of the hands of humans.

When AI crosses this line, it no longer becomes an essential driver of business value. Instead, it works at cross purposes to organizations’ business strategy. This unintentionally counterproductive use of AI is one of the primary reasons the technology is facing so much pushback today.

Where then, is this line? And how can you avoid crossing it?

Defining the value line

The entire reason we are so careful to define toil so precisely is that such a definition gives us a map for what we want AI to automate – and more importantly, what activities we do not want AI to automate for us.

Remember, the definition of toil lists activities that do not create meaningful value. By definition, therefore, activities that do create meaningful value are not toil.

Any organization that believes, therefore, that the objective of its AI transformation is to automate as much work as possible is applying the technology counterproductively, and is thus eventually doomed to fail.

Instead, the goal of any AI transformation initiative should be to maximize customer value while minimizing unnecessary effort (where employees and other users are rolled into the term ‘customer’).

When the AI effort reduces unnecessary effort, it frees up the time and energy of people to focus their time on maximizing customer value.

In other words, the way the AI effort pays off is not to automate value-producing activities, but rather to empower people to spend more of their time and effort on those activities themselves.

Drawing the value line at your organization

To establish the boundary between AI automation and human effort, we can distinguish among three types of work and how we should leverage AI to support them:

  1. Automate toil – when activities take time and otherwise consume resources without creating meaningful customer value, we want to automate them.
  2. Augment value-maximizing human work – we want AI to support human activities that generate customer value without automating those activities.
  3. Support and expand human-critical activities – for activities that depend on uniquely human capabilities including insight, empathy, creativity, human judgment, ethical decision making, and human-to-human connection, AI should play an indirect role only, by freeing up human time for such human-specific value-added activities.

Managers, therefore, must become adept at differentiating these three types of work – keeping in mind that this differentiation work itself requires human judgment and thus falls into the third category.

A simple example: customer service

To illustrate this process, let’s work through a simple example: a customer service rep answering a call and resolving an issue for a customer.

  • The process before AI: after the customer waits in a queue, the rep eventually answers the phone and then works through a flowchart-based script based upon the customer problem, and then either provides a solution or opens a ticket.
  • Inappropriate use of AI: AI answers the phone and provides an automated chatbot-like experience to the human caller, offering whatever solution the AI can come up with (for better or worse) without bothering a real human rep unless absolutely necessary.
  • Appropriate use of AI: AI answers the phone and gathers just enough information from the caller to route the call to the right rep (automation of toil). The rep interacts with the caller (human-to-human connection) while looking at information on their screen that addresses the customer problem or at least gives the rep some alternative solutions to choose from (augmenting value-maximizing human work).

The difference among these three scenarios, of course, is in the customer experience. The first process leaves customers frustrated and angry. The second process (to no one’s surprise) also leaves the customer frustrated and angry – and asking repeatedly for a human representative.

Only the third process, the one where AI supports the human-to-human connection and rapid problem resolution that delights customers, actually maximizes customer value.

The Intellyx take

If the argument in this article gives you a sense of déjà vu, you’re not alone. It is essentially the same argument we made for digital transformation before AI changed the game.

Digital transformation was never about technology. It was always about the customer. The idea was to leverage digital technologies to realign siloed organizations to better meet customer needs – not to disintermediate humans from customer relationships.

Here we are again, making the case that replacing human activities in places where we should augment and support those activities is a bad idea.

Stepping back a minute and looking at the global pushback against AI, whether it be putting people out of work, dumbing down our students, or letting AI agents run rampant, we can address many of the ills of AI by following this simple principle: automate the toil (and only the toil), augment value-maximizing human work, and empower humans to do all the things that humans have always been best suited to do.

If we follow this principle, then maybe – just maybe – AI will live up to its hype.

Copyright © Intellyx BV. Intellyx is the change agent industry analysis and advisory firm focused on enterprise transformation. Covering every angle of enterprise IT from mainframes to artificial intelligence, our broad focus across technologies empowers business executives, IT professionals, and software vendors to leverage disruptive trends to succeed in a dynamic business environment. All of Intellyx’s articles are written by a human, including this one. Image credit: Craiyon.

SHARE THIS: