One of our principles is “Listen before you act” - because we can only develop sustainable solutions if we really understand our customers' challenges. Our proven framework with the four phases Think, Design, Plan, Do and the three dimensions Human, Organization, Technology helps us to shape change holistically.
However, in an increasingly dynamic world, it is not only precision that counts, but also speed. That's why our transformation engineer Alberto has developed an innovative automation solution that uses AI agents to achieve tangible results faster. In this interview, he talks to us about how he developed this solution, the benefits it brings and what inspired him to come up with this idea.
Dear Alberto, thank you for taking the time for our interview, we are curious to know how you proceeded with AI-supported automation. So tell us: what was the idea behind it and what happened next?
Alberto Vera Ramirez: Thank you for inviting me to give an insight into this current and exciting topic at WARGITSCH! The idea was born when the potential of AI, especially ChatGPT, for optimizing consulting processes became clear. At WARGITSCH, we had developed a proven framework over the years, but its practical implementation was often complex. To make this transition easier, we used AI to move from strategy to implementation faster and with higher quality for our clients.
To this end, we developed a minimum viable product (MVP) that tested the first phase of our framework with AI. The aim was to find out how AI-supported work proves itself in practice and how well it can be combined with our existing methods. At the same time, we investigated the technical possibilities, costs and benefits for our customers.
The result: AI not only provided quick insights, but also tangible added value. It increases efficiency and productivity for our customers and at the same time relieves our employees – which creates more space for strategic and creative tasks. The use of AI was therefore a logical next step in the further development and automation of our framework.
That sounds like an exciting approach! How exactly did you use ChatGPT for the implementation?
Alberto Vera Ramirez: In order to integrate ChatGPT into our consulting process in a meaningful way, I worked intensively on prompt engineering – i.e. the optimal design of inputs for the AI. First, I analyzed existing information and prepared it for the first phase of our framework. This was followed by a test phase in which the best formulations and reactions of the AI were determined through targeted trial and error.
In addition, we developed an interaction concept that responds both precisely in terms of content and in the appropriate structure. This resulted in a standard operating procedure (SOP) that defines clear processes for different usage scenarios – including responses in the event of missing information or deviations in the process.
This is how the core of our AI-based approach came about: a sophisticated interplay of input structure, design and interaction – with the prospect of integrating several specialized AI agents in the future.
An important point when integrating AI is, of course, data protection. How do you ensure that confidential information and customer data remain protected and anonymized?
Alberto Vera Ramirez:The protection of sensitive data is a top priority for us. We therefore only use anonymized or fictitious data records in our tests and initial implementations. We have also established internal processes to ensure that no critical information can be passed on to third parties.
In concrete terms, this means, for example, that we have clear guidelines for handling customer data and regularly check which information is actually necessary for the respective use case. When we use real project data, only the absolutely necessary part is fed into our AI platform – and only in secure environments that meet strict data protection standards.
How long did it take you to build this complex process and have you been able to try it out on a real case?
Alberto Vera Ramirez: The development of our MVP took around two months – including conception, information preparation and prototype creation. Despite working on the project in parallel, my prior knowledge of AI and existing material on phase 1 helped me to design the process efficiently.
The prototype was tested in three scenarios during a one-week test phase. The results were promising: around 75-80% of the AI suggestions were directly usable, the rest were reworked. The AI not only provided creative impulses, but also a clear structure – and thus acted as a “right hand” in the process.
Based on these findings, we are now developing the system further – with a focus on optimized bot collaboration, new data sources and advanced prompting strategies – before automating further phases of our framework.
That sounds promising. What's the next step? Will you also implement automation for the other phases and what do you need to bear in mind?
Alberto Vera Ramirez: Our next step is to gradually automate the entire WARGITSCH framework – which is divided into four phases and three dimensions. We are currently focusing on moving the first phase from a beta version to a stable ‘version 1.0’. As soon as this works reliably and we know exactly which requirements need to be taken into account, we will tackle the subsequent phases. At the same time, we are already working on an MVP for our three dimensions to ensure uniform and consistent support here too.
It is important to us not just to develop a large number of isolated bots. Instead, we are planning a holistic ecosystem in which a central ‘master agent’ works together with various specialized agents.
You have a big task ahead of you! How will the automated framework be used by the customer?
Alberto Vera Ramirez: The framework acts like an “X-ray image”: it quickly shows where action is needed and in which phase what needs to be done – whether at the level of people, organization or technology. With the help of AI, we can quickly analyze data, select suitable tools from a pool of over 150 and derive well-founded suggestions. This saves time and costs, increases transparency and enables tailor-made, practical solutions.
In the long term, both the framework and the consulting team are constantly learning – for even more precise analyses and more efficient automation options. The goal is a scalable, AI-supported product that offers real added value – in theory and, above all, in practice.
Thank you, dear Alberto, for the wonderful interview and this deep dive into a complex and exciting topic!
Alberto Vera Ramirez
Fast and effective support is particularly important in economically challenging times.
Get in touch with us – we’ll show you how we can take the pressure off your team with the help of AI!
Your contact person: Alberto Vera Ramirez