The Added Value of AI in Complex Consulting

Before we dive into the topic, let’s first take a look at where and how AI can support in consulting projects. The following image shows a general and simplified workflow of a consulting project. It shows where AI can create added value and what it consists of. As you can see here, for example, during the phase of analyzing a problem, AI can take a role in understanding the problem completely and possibly taking in more variables than a human would. That way, you are able to see the bigger picture by letting the AI take different perspectives. It can also support in describing the problem more thoroughly at a higher quality. In the next sections, we’ll explore more of the value added from AI for certain project steps.

Thinking for yourself makes you smart. Delegating the main aspects of thinking towards an AI makes you lazy. It leads you quickly and uncritically on the wrong track. The temptation is great to radically redesign the consulting workflow and set up an agent framework that only uses humans to nod off the results. I advise against it when it comes to important decisions. Why is that?

1. AI tends towards standard solutions. They are usable, but only average. Without first thinking for yourself, one is willing to confirm such a mediocre solution only because it has been convincingly presented. In my experience, you almost never get creative and truly good solutions out of an AI, even if you are shown many alternatives. For example, I tried unsuccessfully a few times to (re-)create a clever solution I had recognized myself by redesigning the AI prompts. Unfortunately, AI never got to the human solution.
2. It is well known that AI is not always objective, sometimes uses false assumptions and presents them so well that it is only noticeable if you deal with it intensively yourself. Such errors can be found through reviews. However, it is very difficult to find unreasonable limitations of the solution space, as already described under a). All the solutions generated may be correct. But unfortunately, the AI has turned in a certain direction, and you have to notice that first. My experience is that it is very difficult for me to deviate from a solution path that seems to be true or to question it fundamentally, especially if it is part of my own horizon of experience.

That’s why I take the time to think for myself before each sub-step. It’s important that I have my own anchor that prevents me from being drifted away by a strong AI current.

The value of advice for complex decisions lies in the quality of the recommendations. It is quite possible to go through the AI-assisted workflow faster. However, it is worth reinvesting the time gained with the help of AI to improve the quality of the results. These are the AI added values highlighted in blue that can only be created (economically) through the use of AI. You can grasp and think through more content where you previously had to set a termination point due to time constraints.

1. Not (necessarily) generating more output for the customer
   The temptation is great to pour the expanded implementation capacity into more presentation pages. Those who do this usually overwhelm and confuse the customer with the sheer mass instead of convincing him. A crucial value of consulting is to reduce complexity. The aim is to present a previously complex situation in an appropriately simplified way and to show solutions. Decisions usually consist of choosing a single alternative. To achieve this, complexity must be reduced so that the decision point is clearly worked out and the customer is able to make decisions more easily.
2. Leverage customer AI
   If the customer has a functioning internal AI, it can be worthwhile to use it. There are no data protection challenges here, and you can also process entire documents. This is valuable for using internal contextual knowledge. On the other hand, if publicly available AI models are used, it’s a no-go to use any internal information of the client, the company name or complete documents.

In the following, I explain further the added value highlighted blue in the graphic.

AI helps to work out the right problem. After listening carefully to your customer, you can formulate initial hypotheses as to what the real problem is. After that, two steps are recommended:

1. Have the problem description analyzed by AI. It shows if you have missed a point.
2. Let AI take different perspectives to illuminate the problem from a different angle. This way you can find even more challenges and get a more complete picture of the problem. It has proven successful here to give the AI the stakeholders who appear in the problem context as persona and to present the problem from their perspective.

Keep in mind that results produced in this way are to be treated as hypotheses. This does not replace interviews with stakeholders to verify them.

A good problem description is half the solution, if not more. I usually have a conceptual idea of the problem in my head but then I need a lot of time to put it into precise words. AI helps enormously here. It is easy to have formulation variants generated and to create an exact description from them. The clarity of the wording is also a catalyst for understanding the problem.

Both the variety and number of sensible solutions that fit the customer’s business case can be increased thanks to AI. Here, the strength of AI, especially to find standard solutions, is an advantage. The probability of overlooking a major solution path is low. Solutions can also be thought through well by a question-and-answer game in the LLM chatbot. Likewise, you can assign different roles to the AI to work out advantages and disadvantages (e.g. assign the roles of the 6 Thinking Hats method).

The increased number of solutions must be reduced to a number agreed with the customer. AI helps to create estimates for evaluation criteria, if such a method is used for selection, for example, to select three out of ten approaches. Once you have pre-selected the desired three alternatives (with or without customer involvement), AI helps a lot to present the chosen approach in a simple and understandable way.

The AI supports well in estimating the impact of the selected solution. Especially in the risk assessment (SWOT), the AI finds points that would otherwise have been overlooked. It also helps to clearly highlight the key points of the solution so that they are easier to understand.

When working out the chosen solution, AI again helps not to overlook any points. If you initially worked with personae that represent stakeholders, you can now use them again to automatically break down partial solutions to them. This saves a lot of work.

Solutions and implementation are mutually dependent, and I always plan them together. Work packages must be estimated in terms of effort. AI can estimate and justify a reference for this effort. It provides good estimates if you have IT or software-related work packages. But even if the estimate is not quite consistent, I use it as a challenge for my own estimate.

AI can significantly enhance consulting in complex decision-making contexts – but only if it is used as a thinking partner, not a replacement for thinking. Its real strength lies not in speed or automation, but in increasing completeness, precision, perspective diversity, and clarity across the consulting workflow.

Used correctly, AI helps consultants:

• understand problems more comprehensively,
• formulate them more precisely,
• explore solution spaces more broadly,
• reduce complexity more clearly,
• and assess consequences more thoroughly.

However, AI also introduces risks: standard solutions, hidden framing, and subtle limitations of the solution space. These risks can only be mitigated if consultants consciously anchor their own thinking before engaging with AI-generated outputs.

The key takeaway is therefore not “use more AI”, but “use AI deliberately”. In complex consulting, value is created where human judgment, experience, and responsibility remain in control – and AI is used to amplify quality, not to outsource accountability.