Move beyond the Proof of Concept

You have an idea or a proof of concept for something that could have a real positive impact on your business. I’m here to help you to develop and deploy a system that provides the business value you envisioned.

You are dealing with these challenges:

• Are the capabilities of current AI systems even good enough for your use case?
• Does your inhouse team have the competencies to handle this?
  Can you upskill the team or do you need to hire outside expertise?
• Most AI-generated results are great. But how do you deal with the exceptions that are wrong, misleading, or even embarrasing?
• How do you get above the quality bar that you need for positive business impact?
• How do you integrate AI with the rest of your IT landscape?

Why AI-based products are different

1. Investments may become obsolete with the next generation of AI products (GPT-5).
2. Hype can inflate the expectations of stakeholders, leading to disappointment later on.
3. People in your organization are worried about being replaced by AI.
4. A lot more exploration is needed during development and testing.
5. New skills are required for developers, product owners, and QA specialists.
6. Outputs from AI-based systems may not be 100% reproducible.
7. A rigorous evaluation process has to be added on to the usual testing.
8. AI-based systems have weird failure modes and new attack vectors.
9. Operating costs can be much higher than your usual inhouse IT system or SaaS.

How I can help you

"My approach starts from business outcomes, with technology choice being an important but secondary concern. It focuses on iterative development and fast feedback, guided by systematic evaluations. Throughout a project, I emphasize cross-functional collaboration between business and tech teams."

Dr. Matthias Berth
technical product owner for AI-based products

I have hands-on experience in

• Developing LLM-based systems
• Prompt engineering for various NLP (natural langguage processing) tasks
• Systematic evaluation of AI results

Furthermore, I have

• Guided products from proof of concept through to mature market offerings
• Helped teams and organizations to understand the negative impacts of too many concurrent projects
• Established and operationalized key product development metrics
• Educated teams on the benefits of limiting work in progress (WIP)
• Transformed deployments from “all hands on deck” to “business as usual”
• Led algorithm development in image processing and analysis

Over more than 20 years in the software business, I have worked in these roles:

• AI Engineer
• Data Scientist
• Enterprise and Software Architect
• Co-Founder and CTO
• Project Manager
• Software Development Lead
• Product Owner for teams of data scientists, operations researchers, and software developers

Work with me

Here are three ways I can help with your AI projects:

• Hire me as a Temporary Technical Product Owner. I’ll work directly with your teams to guide them in building successful AI applications.
• Book an AI Planning Workshop. Together, we’ll turn your AI ideas into a solid plan. We’ll spot opportunities, identify major challenges, and map out clear next steps.
• Schedule an AI Project Selection Workshop. I’ll help you choose the right AI projects for your business. We’ll decide what to prioritize, whether to build or buy, and how to balance exploration of new ideas with achieving practical results.