AI Trailblazers: Dimitri Rupp Shares Lessons from SCHUHFRIED's AI Journey

In this edition of our AI Trailblazer Series, we sat down with Dimitri Rupp, Head of Operations at SCHUHFRIED, to discuss his AI journey, best practices, and lessons learned from implementing AI tools effectively.

SCHUHFRIED, founded in 1947 in Austria, pioneered the world's first digital psychological test system—the Vienna Test System (VTS). It now facilitates over 13 million test administrations annually across 68 countries, having transformed fields from clinical neuropsychology to transportation safety.

With extensive experience spanning product support, sales consulting, and competitive intelligence, Dimitri shared practical insights on how careful AI configuration and proactive change management can significantly accelerate AI adoption and maximize its ROI.

On AI Implementation and Configuration

What have you found to be most important when implementing AI tools in your organization?

Dimitri: The configuration is extremely important. In Dashworks, the AI control section is crucial because it not only defines what you want and expect from the AI and how it should react, but also what the restrictions or exceptions are.

For example, we initially defined that documents and resources older than 2023 should be less important. However, we later discovered that our sales team needed information from documents dating back to 2018 for comparison with newer versions. When we asked the AI, it responded that it didn't have this information, so we needed to include an exception. The result was a higher quality of answers.

It's very important to implement instructions carefully while also considering exceptions. You don't need to open an entire resource like SharePoint—you can define specific folders and say, "When it comes to questions regarding tests or a specific topic, also include these documents from that source."

On Practical Use Cases

Q: How has your team been using AI tools in day-to-day operations?

Dimitri: In sales, it helps reduce the time for creating specific offers to clients. We can gather information from various sources much faster and get exactly what we need. For specific questions like "Is this test suitable for that kind of customer in this environment?"—the AI provides answers quickly.

In operations, we're very happy with how it helps with onboarding new colleagues. It's like having an additional person by their side. If they have a question and nobody is available, they can just ask the AI and will likely get an answer.

For test services, colleagues can ask questions like "How do I create a new user for our application?" or "What does the login look like?" They type the question and get a structured, step-by-step guide—more beautiful and perfect than I could answer myself.

On Change Management

Q: How do you approach change management and education within the team around AI tools?

Dimitri: Everything starts with someone hearing about AI tools, getting a ChatGPT account, and typing in simple questions. The next level is learning to improve prompting to get better results.

Currently, many people use individual AI tools for research, mind mapping, strategic thinking, and various purposes. The new work technique requires spending time not searching for information but gathering information, comparing different results from different tools, and applying critical thinking. People need to learn this process. When they start working with AI tools, they quickly build confidence, but you need to teach them that it's like having a young colleague who is still learning. You need to review answers, be critical, and sometimes go back and forth to get the complete information.

To help colleagues adapt, I regularly check in with them personally and ask, "How are you doing with the tool? Does it help you? Where are you running into issues?" They don't usually come to me on their own. When I ask my sales colleagues, they might say, "It was great, but lately I had this situation..." and then we can adjust the instructions or how we store information.

In the adaptation phase, it requires a lot of reviewing, going back, asking questions, and helping reduce fear or skepticism. I tell them, "It's just like a student coming to the office on their first day. You need to give more than just a question; you need to provide context." The more context you give, the better the results will be.

On ROI and Measuring Impact

Q: How do you measure the ROI or impact of using AI tools?

Dimitri: We've seen that AI tools like Dashworks can reduce working time by about 68% when used as an enterprise search engine. We have about 15,000 pages in Confluence and over 100 defined technical terms.

Our conversation accuracy has achieved 92%, meaning when sales creates an offer specifically for client needs, the AI helps us think about what we might be missing. We can copy-paste requirements from the client and ask the AI, "Am I missing something? What do I need to think about? Can you analyze these requirements?" This has enabled us to sometimes send one offer with all the information, instructions, and explanations, and have the client respond that it's exactly what they were looking for.

We're also using AI for improving documentation in Confluence, which saves at least 60% of time. Overall, people using these tools become 3-4 times more efficient.

On Security and Data Protection

Q: What concerns have you had around security and data protection?

Dimitri: Data information and data security is a huge topic in Europe, especially with the EU AI Act. Many teams have fears that tools like Microsoft Copilot might have too much access to company information. There's always some mistrust: "If we open this, it's like Pandora's box—it knows everything about our company. Where is this information going? Is it going to Microsoft servers? What are they doing with the information?"

We've assured them that currently, the usage of Copilot does not export any information; everything stays in-house, at least for now. The security center and documentation provided by Dashworks has been excellent, and that's what we need to show to business owners and management to address these concerns.

Advice for Those Starting Their AI Journey

Q: What advice would you give to someone who is where you were six months ago?

Dimitri: First, create a collection of all the help resources for the applications you're using. This helps you avoid going back and forth with support services. Your second major resource is YouTube—watch the latest videos. Also, ask the AI itself: "What are the latest features of this tool I'm thinking of using? What are the advantages?"

Start with simple questions but with the goal of learning about these applications. This combination helps you learn how people are using AI tools in the practical world, and if you have a question, you can ask for specific guidance about configuration or features. This approach will quickly bring you to the point where you're using AI most efficiently.