Hans Honig (Deloitte): ‘The time for AI pilot projects is over’

To what extent is the traditional pyramid model of the Big Four, and specifically Deloitte, changing now that the work of junior staff can be performed much faster by AI?
‘Our entire workforce is undergoing a fundamental transformation, but it is too simplistic to say that AI alone is causing a decline in entry-level hiring. Another major factor is the trend toward global delivery centers: specialized hubs where a great deal of technological expertise is concentrated. In addition, client demand is changing. We are increasingly investing in software – assets – that allows us to complete projects much faster or that provides clients with specific services or insights. As a result, the pyramid has essentially already disappeared. In the future, the organizational model will become even leaner, and seniority will become even more important.
We need experienced people with truly deep technical skills and subject matter expertise – often coupled with industry expertise – to assess the quality of what comes out of AI models. There are, also, significant differences within the organization in how teams are structured. In tax advisory and audit, we as the Big Four still train the majority of the professional group ourselves. In other parts of the consulting firm, such as in tech, the dynamics are very different, and we have more new hires at more senior levels.’

During the 2008 financial crisis, too, almost no juniors were hired. When the economy picked up again a few years later, that turned out to be a problem because the people who should have progressed to mid-level or senior consultants simply were not there. Is that danger not looming again?
‘That is the major strategic question facing us. It is difficult to predict the medium- and long-term consequences for the workforce, while at the same time we need to be making various decisions with a long-term impact. If we get these completely wrong, we could potentially be facing a huge problem in a few years. That is why, when employing people, we should not focus too narrowly on the short term.
Second, it is becoming even more important to not only employ talented people, but to also support their growth so that they stay with us. And third, we need to be convincing in our proposition for lateral entrants, so that we are known as an organization where they can feel at home. In short, we will continue to invest in developing talent across the full breadth of Deloitte.’

Do the people you want to employ need different skills than, for example, five years ago?
‘One of the big questions surrounding AI is how to deploy it in a responsible and reliable way. We use the concept of Human in the Loop, which essentially means that people monitor the decisions being made at all crucial moments in a process. The traditional skill of an accountant – professional skepticism – is therefore more important than ever, and for everyone. Employees must not only have professional knowledge, but above all the ability to critically question the output of AI, to zoom out, and recognize patterns. They must also be aware of the risks inherent in the models regarding ethics and reliability, such as recognizing bias or drift: the gradual loss of accuracy in a model. This requires a professional, critical eye and a deep understanding of how algorithms work. Even though there is an ongoing debate about whether AI can innovate autonomously, I personally see AI agents primarily as a tool and as part of a team, not as a replacement for it. Ultimately, part of the creativity will come from the smart questions we ask, the input we provide, and the data we link to it. The basis should always be the human-tech combination.’

Deloitte has always embraced technology. Does AI help you maintain that leading position? And how do you approach that?
‘We are confident enough to believe it does. When AI was launched, we quickly realized that data reliability is crucial. We developed our own AI platform, Headstart, which is secure and provides access to our own data. It is now used organization-wide. Of course, countless new applications have emerged along the way, and we want to be able to use those as well. Our approach combines top-down management with bottom-up experimentation. For each area of expertise, we have very large workflow technology platforms where people work in a highly standardized way. That is, therefore, top-down. Our agentic AI platform, Solaria, is specifically aimed at experimenting with new applications bottom-up. The trick is to filter everything that comes out of that to determine what really works and adds value, and only then scale it up and deploy it as an agent on those existing technology platforms. That combination works very well. Also, we firmly believe in co-creation with our customers and tech partners. We know a lot, but we do not claim to know everything. The same applies to our customers. We need to discover together what is possible.’

How do you determine when AI applications truly add value?
‘The phase of endless experimentation with pilots is over. More and more companies are approaching the use of AI as a transformation and are focusing on the question of how the technology aligns with the business strategy and can create lasting value. This has truly made it a CEO-level topic. If you want to deploy AI successfully, you have to look through different lenses. How do we get employees on board? How do we structure the organization for this change? What new talent do we need? How do we ensure quality and safety? How do we collaborate with tech partners without becoming completely dependent on them?
In addition to knowledge of the tools themselves, deep sector knowledge is also required. That is one of the reasons behind the recently announced merger of Deloitte Europe, Middle East and Africa into Deloitte EMEA. Our two major European clusters, Central Europe and North and South, have also been merged. This allows us to combine the knowledge from all countries within Europe in, for example, the chemical sector. That simply provides more depth and increases our investment capacity. The strongest argument for bringing the mandate to the European level is a sense of stewardship. We want to pass on a better organization to the next generation and believe this merger is necessary to remain the largest and leading professional service firm in Europe. Geopolitical developments are giving us an extra impetus in this regard.’

Looking at the client side: how important is AI in your services by now?
‘The way we deliver value to clients is fundamentally changing through the integration of AI platforms and smart agents. For example, in strategy consulting, we have an AI platform that helps calculate various scenarios, such as in the areas of customer segmentation or go-to-market strategy. That platform is interactive and is constantly fed with new data, so you get real-time feedback if reality deviates from the assumptions made.’

How does that differ from the traditional consulting you yourself provided ten years ago?
‘First, it results in much greater integration between strategy and implementation.
Whereas you previously had separate evaluation points to check whether a strategy was working, that feedback is now directly integrated into the process, allowing for continuous adjustment. So, from the start, we focus on the full cycle from advice and implementation to the final transformation. We see this change across the board. Another change is the low barrier to innovation. We can build pilots and prototypes very quickly and at relatively low cost, learn from them, and then adjust and scale up. That speed is perhaps the biggest difference compared to the past.’

Are there already examples of applications where AI is realizing its promise?
‘Actually, where it all began, in call centers, finance, the financial sector, our own field. But you see it delivering also in some more physical sectors. For example, we have a client in the maritime sector who uses AI to optimize the logistics of large platforms. There, it is at the core of the business and provides real added value. Within life sciences, it can have a massive impact, both operationally and societally, for example, if you can shorten testing cycles of new drugs with the help of AI. Generally speaking, progress will be particularly rapid in industries where AI processes make it fundamentally easy to implement business models with relatively significant impact. It becomes more difficult in organizations with limited or unreliable data and with a large ‘physical’ human component.’

The revenue model in your industry was traditionally based on hourly billing, with a fixed markup on employee costs. Now you are investing in software that sometimes becomes part of the client’s primary process. Does that change Deloitte’s risk profile?
‘We have already moved away from billing by the hour to some extent, and that trend will continue. The share of our revenue related to technology will increase. This can be through a fixed price that is part of the overall agreement, or clients purchasing those assets separately. We are taking on more risk primarily due to the type of projects we take on. We want to be at the core of our clients' businesses and, in our advisory practice, we are increasingly committing to results or even the value that a transformation delivers. By definition, that means taking on more risk.’

How do you prevent unintended unethical consequences when making AI tools available to clients?
‘We have developed Trustworthy AI, a global framework for ourselves and our clients to ensure that AI solutions are robust, secure, and verifiable from the very first design. This means that every application leaves a clear audit trail, so we can see exactly how decisions are made. A second pillar is the aforementioned concept of Human in the Loop, meaning that a human is always monitoring the process at critical moments.’

How do you convince clients that they can trust your framework?
‘Honestly, simply by being transparent about the risks of AI and explaining how we address them. And, somewhat trite perhaps, but clients can always come to us for an audit of their AI, for a stamp of approval that their system complies with the rules and is ethically sound. I expect the market for AI assurance to grow significantly.’

If you look into your crystal ball, what will AI bring us in the coming years?
‘The most honest answer: the jury is still out. We do not know what is coming, what the second and third-order effects will be, and as yet we cannot even fully grasp the effects of everything that is happening right now. However, I am convinced that it will be huge. You can compare it to the rise of the internet. Initially, expectations were sky-high; at some point, the bubble burst, but in the long run, its impact turned out to be enormous. It will probably be the same with AI. In the short term, there is some overvaluation, and it will most probably be somewhat disappointing; in the long term, AI will bring about major, fundamental changes and – if we approach it sensibly – may well contribute to solving societal problems such as the labor shortage.

This interview was published in Management Scope 04 2026.