Using technology to enhance productivity has long been a goal for many people. Now, hope is with AI. AI will change how we all work. And as with every technology before it, it will help us do more things, do them faster and cheaper. If it doesn’t, the technology will be reduced to a hobby.
But exactly when are we better off with more things, faster and cheaper? Increased productivity has always been a goal for economies and countries. But not all work a company does benefits from increased productivity – it can be quite the opposite. With the current AI expectations, we will see a lot of missed opportunities and misdirected investments when not enough attention is given to the nature of work itself and the people doing it.
In this article, we dive into the complex dynamics of productivity, exploring its diverse aspects within operational and knowledge work. By understanding the nuances of individualistic and collective views of work and the influence of corporate culture, we shed light on how AI can effectively empower productivity.
Productivity is the rate at which a company or country makes goods, usually based on the number of people and materials necessary to produce the goods. One of the first things taught in economics is that increased productivity is good. It’s good for the economy, so it makes everyone better off. Notably, it’s good, on average, across the entire population.
But while this is true when summed up at a macro level, it’s not the same for every single company in the economy, where productivity is a mixture of effectiveness and efficiency in the context of risks and rewards. While productivity is simply the correlation between input and output, efficiency denotes doing things resourcefully. Effectiveness is doing the right things.
For a company to effectively leverage AI or any technology, it’s crucial to pay attention to these differences in combination with how we think about work and the nature of the work itself.
Individualistic and collective views of work
We are all different and unique, so we will think differently about work, even when doing the same job. Regarding work, there are two contradictory viewpoints to consider: the individual and the collective.
Many people view their work from the perspective of a task list. Work is functional, typically expressed as “my job is to do…” – this is where the level of performance and efficiencies are thought of as individual and measured in relation to one’s personal goals and targets.
The opposite viewpoint is where the work results are external to the execution of the work itself – where an individual would say, “My job is to provide for…” and view work as a collective undertaking towards a common goal.
Without going into too much depth or stereotypes, a corporate culture reflects the beliefs and behaviors that determine how a company’s employees and management interact. It’s influenced by cultures and traditions as well as trends and practices. But overall, the generalized behaviors and culture will be dominated by either the individualistic or the collective view of work.
The nature of work
For most of the last century, work was divided based on the people who performed it. Until the undeniable success of the Toyota Production System, work was classified as blue-collar, manual labor, or white-collar, professional work.
Since the introduction of Lean and the Theory of Constraints, we now have a more meaningful understanding of how work flows through an organization before the results reach its customers.
Some work is transactional; it can be completed and evaluated in a single transaction. Most modern work, however, requires a sequence of tasks by multiple people to be completed before the output reaches customers. It’s process work where the result is not immediately evident.
From the flow perspective, distinguishing between operational and knowledge types of work goes deeper than education levels and shirt colors.
Operational work is simple and repetitive by nature. But most significantly, the output, or result of the work, is not unique. As the output is repetitive, effort estimations can be precise, with statistical analysis and quality determined against set standards.
Operational work can be found in services, transaction processing, accounting, and of course, in manufacturing. Where operational work is transactional, most can be easily automated by physical or software robots. For example, a chef still cooks the meal you ordered, while AI is already doing most of a Border AAgent’swork validating your identity and passport.
Knowledge work, on the other hand, is creative work. It typically requires decision-making and often collaboration around complex tasks. The key characteristic of knowledge work is that the output is a one-off, even if the activities are repetitive. When the output is unique, quality cannot be determined upfront; results need to be determined based on feedback.
For example, when the border agent determines if your reasons for entry are sufficiently legitimate, the conclusion and resulting decision applies to you only, i.e., it’s transactional knowledge work. When the agent joins a task force to prevent illegal border crossings, it’s an example of process-driven knowledge work.
Below is an overview of four scenarios our chef will likely encounter in their business.
Given how work itself can be different and the differences in our approaches to it, there should not be a surprise that productivity plays a different role in each scenario.
Productivity for operational work
For operational work, efficiency is essential as the work is repetitive. Efficiencies are linked to results. Given timeframes and quality are kept on the same level, more and faster is better – in most cases.
If the work can be completed in one transaction, individual productivity goals can be added up to total productivity. The more passports every border agent checks per hour, the shorter the queues will be.
For a process, i.e., work that requires a series of tasks to be completed, individual efficiencies only add up to total productivity when it happens to be at the bottleneck. Increased work rates at any other point than the bottleneck is, at best, ineffective. Often, it will lead to longer waiting times and excess inventory.
This is a particularly important point: whether using technologies or not, the amount of final output will only increase when the throughput of the constrained resource is increased. If a chef increases their productivity to 100 daily specials per hour, but the waiting staff still only serve 80 diners per hour, the increase will create waste or at least waiting time.
The effect of corporate culture
From the example above, a company with a culture based on individual goals and rewards is much more at risk of suboptimizing. Identifying and resolving common challenges is much more natural in companies where the norm is to work collaboratively towards common goals.
Productivity for knowledge work
In knowledge work, efficiencies are less significant than effectiveness. Since the output of the work is unique, it’s the effect of the output that needs to be measured and evaluated first, not the efficient use of resources going into it.
For transactional work, e.g., for a chef composing a new menu, the risk of being inefficient is relatively low compared to the risk of being ineffective. It is far more important that the menu is popular with the patrons than that it is efficiently composed.
The effect of corporate culture
Again, a corporate culture where goals are external will be better at listening to feedback. This is doubly important for knowledge work completed through many process steps.
When our chef wants to produce his own cooking show, many pieces need to fall in place. Not only is the work collaborative, but it also depends on expertise from many different individuals. Having a common goal and a collective view of the work will lead to better results than a scenario where everyone is focused on meeting their targets. From the example above, a company with a culture based on individual goals and rewards is much more at risk of suboptimizing. Identifying and resolving common challenges is much more natural in companies where the norm is to work collaboratively towards common goals.
Our guide for applying AI to knowledge work
Productivity is more complex when it comes to knowledge work. Without consideration of the prevailing corporate culture and the flow of work, introducing AI technologies will likely not live up to expectations.
Seek effectiveness before efficiencies
When introducing technologies to increase the productivity of knowledge work, consider the risks and uncertainty around effectiveness first.
Focus on outcomes
Ensure that goals are external, not only to yourself, but external to your team, and preferably to your company. Follow the principle of delivering value early and often, using the technologies and ways of working that enable incremental delivery.
Identify and exploit bottlenecks
Map the flow of work through the organization to the impact on customers’ or employees’ experiences and optimize the flow of work end-to-end. Identify the activities that constrain the total throughput and invest in increasing efficiencies only at these points.
Increase the rate of feedback
The best strategy to minimize uncertainty is through testing and learning. When work is complex, our guiding principle is to ensure quality through fast feedback.
It is most productive to complete work instantly, with minimal resources. However, the economic value comes not from efficiency alone but from the combination of effectiveness and efficiency.
A company with a corporate culture that rewards collective performance aligned to its customer experiences will have the upper hand in leveraging AI and technologies that promise to increase productivity. Especially for knowledge work, where any changes in productivity that do not improve effectiveness are likely wasteful, regardless of being caused by an increase or decrease in the rate of output.
If you consider implementing AI solutions in a culture where individual goals and rewards dominate over collaboration and external objectives, this is a real risk.
I have written a few articles before this one, but this one is unique and results from knowledge work. How productive was I in writing it? How much aid was provided by AI in the process? The answer to both questions is a little, but that’s irrelevant. The all-important outcome is the effect it has on readers.
If you enjoyed the reading, visit our Insights page, where our team explores and shares insights on topics such as How to add more intelligence to your application with Machine Learning, It’s not the people it’s the system and our thought paper Intelligent Automation of knowledge work.
Definitions used: (from Cambridge dictionary)
Bottleneck – A problem that delays a process or stops it from continuing.
Corporate Culture – The beliefs and ideas that a company has and the way in which they affect how it does business and how its employees behave.
Efficiency – The quality of achieving the largest amount of useful work using as little energy, fuel, effort, etc. as possible.
Effectiveness – The quality of being successful in achieving what is wanted.
Productivity – The rate at which a company or country does useful work.