The Three-Bucket Method

Image generated with Artlist under licence.

The evolution of artificial intelligence is now on the agenda of a growing number of companies and, consequently, of the professionals who work for them — or are considering joining them.

However, the sheer volume of information we're exposed to is overwhelming, sometimes even exhausting. It feels as though the world we know could disappear within a matter of months.

In the business and professional world, this uncertainty is creating confusion and leading to some rather random responses, to say the least:

• Those who have rushed straight into implementing technology, cutting labour costs and expecting everything else to magically fall into place.

• Those who continue to believe that AI will have little impact on their industry, particularly in sectors that rely heavily on manual or on-site work.

• Those who are conducting experiments in specific areas to assess the technology's potential and implications.

• Those who have developed a strategic plan, established ethical guidelines, trained their teams and begun implementing AI.

• …

  
  

There are almost as many approaches as there are companies. And the same applies to professionals and employees:

• Those who experiment with every new tool that appears.

• Those who have simply dabbled in one or two of the best-known solutions.

• Those who, having tried them, believe AI's capabilities and potential contributions are still very limited.

• Those who haven't even started experimenting, apart from perhaps trying something in their personal lives.

• …

  
  

What is clear is that generative AI is a disruptive technology evolving at extraordinary speed, with capabilities that continue to expand. We therefore need to accept that AI is here to stay and, more importantly, learn how to make it work to our advantage.

That said, we need to be realistic: not all industries or professional roles are currently being affected in the same way.

That's why, as the saying goes, we need to move forward "slowly but surely." Every company needs to understand how this technology is reshaping its industry and identify where it currently stands on the technology adoption S-curve.

Image generated with Artlist under licence.

This curve illustrates the different stages a technology goes through as it becomes established. We could also use Gartner's Hype Cycle, although that focuses less on actual adoption within an industry and more on assessing whether a technology is becoming established, experiencing intense hype or falling into the "trough of disillusionment" after a promising start.

The S-curve can be divided into the following stages:

S1. Introduction Stage

Adoption remains low — below 10% — and is primarily driven by the most innovative companies.

The technology generates curiosity but also scepticism. The first to experiment with it are organisations and individuals with a strong technological mindset and a willingness to try new things.

At this stage, pilot projects tend to be isolated, with no clearly defined processes or measurable return on investment (ROI).

S2. Take-off Stage

This is the early adoption stage, with adoption still below 40%. The technology is approaching what is commonly known as the tipping point.

Real success stories begin to emerge publicly, and the productivity gains achieved by AI users become increasingly evident. This stage is led by industry leaders and companies seeking to differentiate themselves.

Within organisations, internal policies begin to take shape, corporate licences are purchased and formal employee training programmes are introduced.

S3. Rapid Growth Stage

Adoption reaches the majority of the market, ranging from 40% to 80%.

AI stops being a distinctive competitive advantage and becomes a basic requirement. Most companies in the industry are already using it because they don't want to fall behind.

AI tools become directly integrated into everyday software, including word processors, video editors and management systems.

S4. Maturity and Stabilisation Stage

Late adopters come on board, and adoption reaches between 80% and 100%.

The technology becomes invisible. People no longer talk about "doing a task with AI"; the task simply gets done. This is when the most traditional industries and companies finally adopt the technology, along with those who have been most resistant to change.

The market stabilises, and innovation begins shifting towards the next emerging technology

Based on these criteria, we need to identify where we currently stand. In addition to observing market signals, the following guidelines may help:

• If the tools we use every day already include AI by default, we're at least in a stage of mainstream adoption.

• If an industry has numerous specialised courses and guidelines on using AI, adoption has already taken off.

• If the main debate revolves around ethics or job losses, the industry is probably experiencing a period of rapid adoption growth.

  
  

Once we've established, at least approximately, where our industry stands, it's time to look inward and analyse how AI might specifically affect us and the people we work with.

This is where the three-bucket method comes into play.

Since the Industrial Revolution, we've lived in a society that has rewarded productivity and efficiency more than creativity and out-of-the-box thinking: "Just get it done, and don't think too much about it."

As technology has advanced, it has gradually taken over the most basic productive tasks or those offering the greatest opportunities for efficiency gains.

Traditionally, this mainly affected lower-skilled workers and so-called blue-collar jobs. However, generative AI represents a white-collar revolution. It doesn't just affect physical tasks — where machines, many already powered by AI, have long been used — but also those who believed they were safe because they worked in front of a computer or had earned a university degree and a master's.

Add to this the fact that we spend very little time thinking about how to improve what we do and far too much time trying to do it faster, and we have a serious problem.

The Three-Bucket Method

Image generated with Artlist under licence.

The three-bucket method is designed to help us learn to work and coexist with AI through a reflection process divided into three stages:

BUCKET 1

Which activities that we currently perform ourselves could AI carry out autonomously, delivering results of equal or better quality?

BUCKET 2

Which activities produce better results or higher-quality work when we collaborate with AI?

BUCKET 3

Which activities can only be performed by humans because AI cannot carry them out independently, and its assistance doesn't improve the outcome?

These are three simple questions, but they require careful reflection. The answers will depend largely on our knowledge, responsibilities and working environment.

If more than 50% of our activities fall into Bucket 1, we're facing a clear risk. A significant part of our role can already be performed by AI, and the proportion of automatable tasks is likely to increase over time.

If the proportion of activities in Bucket 1 is below 50%, we might assume we're safe. But that's not necessarily the case. Our industry or company may simply not have reached an advanced stage of AI adoption yet.

This means we have a little more time, but we can't afford to become complacent. We need to keep a close eye on how the technology and our environment evolve.

If more than 50% of our activities fall into Bucket 2, it's a sign that we're learning to work alongside AI and make the most of its capabilities. Even so, we shouldn't become overconfident, because some of these activities may naturally migrate into the first bucket as technology advances.

Nevertheless, this situation gives us time and frees up resources to strengthen Bucket 3, which encompasses the capabilities that make us unique.

There are very few professions in which Bucket 3 can be considered clearly dominant. However, it should become our guiding principle for "mastering AI": developing the ability to contribute vision, judgement and purpose, and then using technology to enhance what we want to achieve and make the process more efficient.

The three-bucket method can be applied at an individual, team, departmental or company-wide level. However, the results will vary significantly depending on how each person, department or organisation approaches the adoption of a technology such as AI.

In other words, we can increase or reduce the risks to our professional future depending on how we respond to it.

As a general rule, extremes are rarely a good idea: neither believing that human teams will be replaced by machines almost overnight nor resisting change and insisting on working as though we were still in the 20th century.

The key is to carefully analyse how AI can be integrated into our everyday work.

How can we make it a valuable tool for our teams? How should we train them? Which human capabilities do we need to strengthen?

That's where both the challenge and the risk lie.