The Path of dkd’s AI Transformation

About adventurers, stabilisers and balancers

Almost no one can escape it – whether as individuals, as a team or as an organisation. We’re all caught up in what is known as the AI transformation. Some describe certain aspects of AI as a ‘gamble’ or the ‘Wild West’. The unknown breeds uncertainty. Do we need sheriffs and marshals? Where is the ranger with his compass?

It is not the individual AI technology that determines the success of the transformation, but how an organisation deals with change. Companies that are in direct competition with one another are introducing AI because they expect it to improve their internal processes, workflows and, above all, efficiency. The focus is usually on improving performance through more efficient day-to-day working. But what does such a transformation require, and how do you prevent a technology from being introduced simply because it is available – rather than because it serves a clear objective? How do people within a company remain mentally resilient during an AI transformation?

As an IT company, we naturally asked ourselves the same questions. This article details our journey through AI transformation – we would be delighted to support you on your own journey with our experience.

Mindsets

An AI transformation is not a short sprint, but a long expedition fraught with many unknowns. As with any good journey, you need different types of people in the team to reach milestones steadily. When an organisation begins to transform itself through artificial intelligence, very little of what is on the slides of a kick-off presentation actually comes to pass. For many things, there is no straight path from A to B, no homogeneous group marching in unison towards a new AI future. Instead – almost inevitably – different attitudes towards the new emerge. In my view, these attitudes can be broadly categorised into groups when it comes to AI transformation (and this applies not only to AI-related issues):

Adventurers, Stabilisers and Balancers

None of these groups is better or worse than the others. They are all necessary for change to be possible. Clearly, nobody is just one type. No matter which conceptual framework we use, people are always a mix of many characteristics. In this model, that means we are all adventurers, stabilisers and balancers, just to varying degrees. Such conceptual frameworks help us to better understand obstacles and conflicts so that we can derive solutions from them. Why do we behave counterproductively towards one another in meetings? Why do we find it difficult to agree on shared priorities and goals? What do our roles have to do with our behaviour? How can we better support one another despite our differences?

One important group is missing. I’ll discuss the fourth group in an organisational transformation at the end.

The Adventurers

The Adventurers are the ones who don’t wait to understand new tools before trying them out – they try them out in order to understand them. No sooner is an AI tool available than they’ve already tested it in three different workflows, explored its limits, refined the prompts and enthusiastically told their first colleague all about it. They’re driven by curiosity, coupled with a high tolerance for uncertainty and failure. A prompt that doesn’t work isn’t a setback for them, but rather a lesson for the next iteration. This attitude is worth its weight in gold in any transformation. Adventurers generate the first visible successes; they test where others still hesitate; and they provide the stories that make an abstract strategy tangible. Without them, individual steps in an AI transformation would probably only move beyond the theoretical stage at a late stage – if at all. The flip side is just as real: speed can outpace diligence. What is an exciting experiment for the adventurer may be a risk for other parts of the organisation – one that has not yet been properly assessed and tested. Legitimate questions – such as those relating to IT security, IT architectures, data protection and product maturation – will normally always arise. Adventurers need freedom and an environment that supports their energy, but they also need the stabilisers.

The Stabilisers

The Stabilisers ask the question that is often overlooked in the initial euphoria: what happens if it goes wrong? They are not against change – they are against change without a solid foundation. Their focus is on existing processes, reliable quality, legal and ethical guidelines, and the trust of customers, which has been built up over many years and could be damaged in a moment of carelessness. This scepticism is often misunderstood as a ‘brake’ in transformation processes. In fact, it is usually the opposite: a prerequisite for change to remain sustainable in the first place. Stabilisers ask about data protection before a tool is put into production. They insist on review processes when AI-generated content reaches customers. They remind us that a tool which impresses internally must first prove externally that it delivers on its promises.
Without stabilisers, every transformation would eventually stumble at the very point where speed meets reality – whether through a data breach, faulty AI output reaching the customer, or a breach of trust that cannot be repaired as easily as a prompt.

The Balancers

Between the Adventurers and the Stabilisers, there is a third group that is rarely vocal, but often reduces or even eliminates conflicts and obstacles: the Balancers. They act as mediators and find it easy to recognise different perspectives. They listen to the enthusiasm of the Adventurers and the concerns of the Stabilisers – and actively seek a way to reconcile the two. Not through half-hearted compromises, but by asking: How can we make use of the new without losing what has been tried and tested?
Balancers are the ones who turn an idea into a process with clear guidelines. They understand that speed and care need not be a contradiction, but can be approached sequentially or in parallel. In the dkd AI transformation, it is often precisely this quality that mediates and ensures that friction turns into progress rather than stagnation.

All three are needed

An organisation consisting mainly of adventurers changes rapidly – and in doing so risks squandering the trust it actually set out to strengthen. An organisation consisting mainly of stabilisers remains safe – and eventually becomes irrelevant because the change around it is happening faster than its own. It is this healthy interplay that defines an organisation’s transformation: bold enough to take risks, careful enough to maintain trust, and wise enough to bring the two together. 
At dkd, this is evident on both a small and a large scale. In teams where one person tries out a new tool, another asks about the data protection implications, and a third person uses this to formulate a clear internal standard. And on a larger scale, when individual experiments with LiteLLM, Open WebUI, n8n, PR-Agent or the local AI testbed coalesce into a shared AI strategy – one that wasn’t conceived on the drawing board, but emerged precisely from this interplay of risk-taking, scrutiny and mediation.

The framework that holds everything together: Agile thinking & working

It strikes me that the path of agile transformation embarked upon by dkd almost 20 years ago was primarily intended to enable us to master the AI transformation today. Thanks to agile thinking, we as an organisation have every opportunity. It seems so self-evident, yet it is not. Over more than 20 years, we have worked hard to develop this approach and have deeply internalised it. Our clients also benefit from our knowledge. And yes, of course we too have experienced our share of challenges and setbacks. We have grown through them. Agile thinking and working have always helped us to succeed in every mission, and this remains true now as we navigate the AI transformation.

There are many agile frameworks and training courses on the market – perhaps too many – and it’s easy to lose sight of what really matters in agile working. Five principles serve as our compass, helping us to find our bearings time and again and set out on the right path:

Transparency. Execution. Review. Adaptation. Simplicity.

That sounds temptingly simple, and we’ve certainly heard such terms used very often. Nevertheless, agile working repeatedly presents challenges, such as the problem of cognitive biases (where new information is interpreted through the lens of old knowledge). Have we truly internalised agile working, or have we – and do we continue to – explain and practise agile terms and methods using tried-and-tested approaches? Are we simply doing the old ways with new terminology?

• Transparency means that experiments and topics are easily accessible and visible to everyone, rather than disappearing into one-to-one chats, private test environments or unspoken agreements. Anyone testing a new model, building an automation or refining a prompt should do so in a way that allows colleagues to find out about it – for example, via tickets, reviews and retrospectives, through open lines of communication and meetings, via OKRs, lists, Kanban and news updates. Only in this way does a multitude of individual attempts coalesce into an overall picture, fostering shared understanding and collective thinking within an organisation. Creating transparency therefore requires a great deal of willingness on the part of everyone involved.
   
• Execution or taking action means making a conscious decision not to talk ideas to death before they have even been tried out. Instead of months of planning, this leads to a prototype, a pilot project, a small-scale trial – something from which real experience can be gained, rather than simply making assumptions. This value gives the adventurers the space they need, but also requires everyone else to be willing to accept an unfinished result for what it is: an interim outcome, not a finished product.
   
• Review is the moment when what has been done turns into insight. This is where the stabilisers raise their legitimate questions: Does it work reliably? Does it live up to what a customer would expect from it? At dkd, verification does not take place as a one-off gate check at the end, but regularly and iteratively – in sprint retrospectives, in reviews, and through open feedback from the teams. It is the point where enthusiasm meets reality, and that is precisely what makes it so valuable: without it, every experiment would remain an experiment forever, without ever becoming a reliable tool.
   

• Adaptation is the outcome of the review. What has proved viable is expanded, integrated into existing processes and provided with clearer guidelines. What has not worked is changed or discarded – not as a failure, but as a necessary part of the journey. Adaptation is the ability to continually readjust one’s course without completely losing sight of the direction.

• Simplicity, finally, is the principle that prevents four good principles from turning into an overly complex process. At dkd, this does not mean that topics are dealt with superficially, but quite the opposite: it is a conscious decision to introduce only as much structure, as many tools and as many rules as are actually needed. This principle applies equally to all three groups: it reins in the adventurers when their enthusiasm for experimentation threatens to lead to a proliferation of isolated solutions. It reins in the stabilisers when justified caution turns into an over-regulated process. And it serves as a compass for the balancers, who constantly ask: ‘Is there a simpler way to do this?’

These five principles operate in a constantly repeating cycle. Action is always the starting point and needs to be made visible. Transparency enables review; review leads to adaptation; adaptation in turn creates new scope for action – and simplicity ensures that this cycle does not itself become a burden. This framework has not remained an abstract methodology, but has been our practice in client and internal projects for many years. Specific areas of expertise have emerged from this, including in the course of the AI transformation. We showcase our products and services in further blog articles and on the dkd AI Expertise page.

Those Who Wait

And finally, I shall explain what the fourth group is all about.
In any major organisational transformation, there are also those who wait, who do not really belong to any of the groups mentioned above. Waiting is not, in itself, a negative thing and can have many reasons behind it. Here, there is a desire to be taken into account. Ensuring that these individuals are taken into account must also be a success in a transformation and should always be kept in mind. This can be achieved effectively through the five principles mentioned above. In this way, one can prevent misinterpretations, misunderstandings, rumours and conflicts, particularly during phases when the ‘Adventurers’ and the ‘Stabilisers’ are not cooperating so smoothly. And this is precisely where experienced ‘Balancers’ are needed – people who can recognise the causes of negative sentiments and conflicts in good time and counteract them.

We are right in the thick of the dkd AI transformation, and the journey continues.
We are happy to share our knowledge and support you on your path to AI transformation.