AI from productivity to P&L: nothing happens by chance

Since AI became a mass phenomenon, the quest for productivity has become something of a holy grail. We expect it to save time, speed up tasks, and process larger volumes more quickly, with the underlying idea that all of this will ultimately generate value. Indeed, the reasoning seems obvious: if we work faster and do more with the same amount of effort, the bottom line should follow. However, this link is far from automatic.

This confusion is not new and has been seen with every technological wave. What is changing with AI is the scale of expectations and the speed with which we expect it to bear fruit, all because its effects appear fast, sometimes even before the business has asked itself the question that matters to senior management: where, in concrete terms, should this productivity be reflected in the P&L?

This brings us back to the eternal and misunderstood debate about what we can expect from technology (Technologies sell productivity, but businesses want revenue). With all due respect to those who swear by it, productivity does not bring much in itself unless we decide what to do with it, which already assumes that we will achieve productivity gains, which is not always obvious, at least in the early stages.

In short:

• AI is seen as a driver of productivity, but this productivity does not automatically translate into financial value without clear choices about its use.
• The introduction of AI also generates costs (supervision, control, quality) that can temporarily degrade results until its uses are stabilized.
• Value creation depends on explicit governance: without arbitration on the use of time or gains, these are absorbed with no visible effect on the P&L.
• Increasing production capacity through AI only creates value if demand follows; otherwise, the effort is useless or even counterproductive.
• Without a strategy to transform gains into concrete value, the business ends up reducing costs (often human) to justify the expected effects on results.

AI creates opportunities, but it also creates costs

There is a lot of talk about what AI can achieve, but less about what it adds and costs in various ways. However, AI introduces new costs, and not just technical costs. There are the tools, of course, but above all there is everything that surrounds their use: supervision, validation, corrections, and the need for increased vigilance.

Today, AI produces errors, introduces variability that teams must absorb, and creates a constant need for control, which only increases the famous “work about work” (Work about work: when the reality of work consists of making things that don’t work work). As a result, the cost of quality often increases initially. This is a fact, and to deny it is to ignore what teams are actually doing to get the job done despite the inconveniences caused by the introduction of new technology and the potential transformation of work that it brings.

This point has a significant impact on many things. Until usage patterns stabilize and work processes are adapted, AI can temporarily negatively impact the P&L. The value is neither immediate nor guaranteed. On the contrary, it is contingent on many factors, almost always delayed, and depends on how the organization navigates this initial phase or even anticipates it.

The P&L only changes if we decide where the value should appear

Once we accept that AI creates both opportunities for improvement and costs, we must ask ourselves how all this can be reflected in the P&L. When value is actually created, it can be reinvested in various ways, but we still need to consider the trade-offs (Without governance, the gains from AI are virtual).

This is somewhat similar to Parkinson’s law. Very often, productivity gains are absorbed by employees for their own benefit to improve their working conditions (Are you familiar with Parkinson’s law on how your employees manage their own time and productivity?) and in any case, a recurring problem is that employees don’t know what to do with the time that AI saves them (How Is Your Team Spending the Time Saved by Gen AI).

In short, if the business does not think about how to reinvest the time saved, employees, managers, and the organization as a whole will not be short of ideas for reinvesting it, sometimes for the better, sometimes for the worse.

In other words, it is not AI that creates value, but it is up to the business to decide what to do with it.

Demand is not infinite, and capacity is not always the issue

It is also important to remember one obvious fact: increasing capacity only makes sense if this additional capacity can be absorbed. However, many businesses do not have a supply problem, but rather a problem with outlets, prioritization, or distribution. In these situations, producing more does not create any additional value and may even result in a loss of money, particularly in the service sector and not only for the self-employed (For freelancers, productivity does not always pay off).

But as long as we are talking about productivity, we still need to know how to measure it, not only once AI has been deployed, but also and above all before, in order to have a basis for comparison, which is never done (How can we measure the productivity gains of AI?).

Transforming potential into tangible value

As long as a use case remains experimental, it will not produce anything sustainable, just as a practice cannot support the P&L unless it is integrated into the business (Stabilize to move forward: why experimentation alone does not produce results with AI).

This observation is in line with certain recent analyses whose logic is clear: value does not come from experimentation, but from the path that transforms experimentation into industrialized and sustainable practice (Mapping AI Value Pathways).

But stabilizing does not mean freezing. It means deciding what should remain of the existing, what deserves to be integrated, and what can be abandoned. Without this, the potential created by AI remains latent.

When you don’t prepare for P, you end up adjusting L

There is one point that is often avoided, even though it is central to P&L logic. When a business does not prepare to transform the gains from AI into products or value creation, they remain invisible economically, but expectations do not disappear. This must be reflected in the only figures that interest executives and shareholders.

In other words, AI has an impact on intermediate indicators, but this impact is not (yet) visible in the books.

Unable to channel this impact towards P (read above), the business almost always ends up acting on L. Not because of a well-thought-out strategy, but precisely because of a lack of strategy and because it is the only lever that is immediately visible in the income statement.

When, one day, we find ourselves with productivity gains that we do not know how to monetize or value in one way or another, we have no choice but to cut costs, particularly salaries and jobs.

Bottom line

AI does not create financial value on its own. It creates potential and expands the realm of possibilities, but this leads nowhere unless the business decides where it wants to go. Value is always the result of choices about how to use gains, allocate resources, accept the associated costs, and decide when a practice is mature enough to be integrated into work design and workflows.

For senior management, the message is simple: if the P&L is not changing, it is not because AI is not working, but because the business has not yet decided how to transform what it has gained into results and/or has not prepared for it.

To answer your questions…

Image credit: Image generated by artificial intelligence via ChatGPT (OpenAI)

Related articles:

Two interpretations of AI in the workplace: agentic uses and capability overhang Adoption and impact of AI: lessons (and limitations) from the latest McKinsey and BCG studies Who is handling your artificial intelligence projects? Probably not the right people. Without governance, the gains from AI are virtual Generative AI in the workplace: revolution or illusion? Technologies sell productivity, but businesses want revenue