As data governance grows in importance, businesses need to take control of their data to avoid chaos. In this interview, we talk to Kateřina Ščavnická, a data governance expert at Kiwi.com! Kateřina also speaks at the Data Innovation Summit 2025. With experience in a fast-moving company, she has seen how quickly data chaos can spiral out of control.
Kateřina shares insights into how organizations can prevent common governance mistakes and stresses the importance of embedding governance practices early in the data lifecycle. She discusses how tools like data contracts and semantic layers play a critical role in maintaining consistency across systems. She also highlights how AI-driven technologies are transforming data governance from a reactive task to a proactive approach.
Get ready for advice on how to integrate effective data governance practices into your organization. Discover how to transform the way you manage and trust your data!
Hyperight: Kateřina, what inspired you to focus on data governance, and how has your journey at Kiwi.com shaped your perspective on this topic?
Kateřina Ščavnická: Working in a fast-moving company like Kiwi.com, I’ve seen firsthand how quickly data chaos can spiral out of control. Every shortcut and quick fix adds up, making it harder for teams to trust and use data effectively. I became passionate about data governance. I realized it’s not about control—it’s about empowering engineers to build reliable, scalable data systems. The more we approach governance as engineering rather than firefighting, the better our data (and our sanity) will be!
Hyperight: How do you define ‘data debt,’ and how does it impact the productivity and efficiency of a team or company?
Kateřina Ščavnická: Data debt is what happens when short-term workarounds pile up, creating long-term problems for the company. It leads to unreliable data, constant troubleshooting, and wasted time fixing broken reports instead of building new things. The key is stopping data chaos at the source rather than cleaning up the mess afterward.
Hyperight: In your experience, what have been some common pitfalls companies face when dealing with data chaos, and how can they be avoided?
Kateřina Ščavnická: The biggest mistake is treating data governance as an afterthought—something to fix later instead of building it in from the start. Another pitfall is relying too much on manual fixes and interactions instead of designing systems that prevent problems in the first place. Companies can avoid this by using tools like data contracts and semantic layers. The goal is to enforce consistency by embedding governance directly into engineering workflows. Small, practical governance steps can make a huge difference!
Hyperight: Can you explain how data contracts and semantic layers can help prevent the messiness that leads to data chaos?
Kateřina Ščavnická: Data contracts set clear expectations between data producers and consumers, ensuring that schema changes don’t break downstream systems. They create a structured, predictable data flow, reducing the risk of unexpected surprises.
Semantic layers, on the other hand, provide a shared definition of business metrics, preventing discrepancies across teams. Together, these approaches help companies stop firefighting bad data and instead build reliable, scalable systems.
Hyperight: In your presentation at the Data Innovation Summit 2025, you talk about the importance of making governance practical for engineers. What does that look like in a real-world scenario?
Kateřina Ščavnická: Engineers hate bureaucracy, so governance needs to feel like a natural part of their workflow. This means automating documentation, integrating governance checks into CI/CD pipelines, and using tools engineers are already comfortable with. AI is playing a bigger role here, helping teams detect inconsistencies, predict data quality issues, and even auto-generate documentation.
In the future, I see AI-driven governance tools acting as co-pilots for engineers. They will proactively guide them toward better data practices instead of just enforcing rules after the fact. The goal is to make governance seamless—something that helps engineers rather than slows them down.
Hyperight: What are some of the first steps an organization can take to shift from firefighting data issues to proactively managing data governance?
Kateřina Ščavnická: Start by identifying the biggest sources of data chaos—whether it’s unclear ownership, unreliable data pipelines, or inconsistent metrics. Then, implement small but impactful changes, like introducing data contracts to standardize inputs and setting up automated quality checks. AI-powered anomaly detection can also help teams identify data issues before they cause major disruptions. The shift happens when governance moves from being reactive to proactive, where teams can predict and prevent issues instead of constantly fixing them. The future of governance is intelligent, and organizations that embrace AI-driven solutions will have a huge advantage.
Hyperight: What challenges do engineers typically face when implementing data governance practices, and how can they overcome them?
Kateřina Ščavnická: The main challenge is that governance is often seen as an annoying, top-down mandate rather than something that actually helps engineers. Another issue is that many governance tools feel disconnected from day-to-day engineering work. The solution is to embed governance into existing workflows—things like automating schema validation, integrating quality checks into CI/CD, and making documentation effortless. If governance feels practical and engineer-friendly, adoption happens naturally.
Hyperight: How do you see the future of data governance evolving, with emerging technologies like AI and machine learning?
Kateřina Ščavnická: AI and machine learning are transforming data governance from a reactive process into an intelligent, self-improving system. We’re already seeing AI-powered data lineage tracking, anomaly detection, and automated metadata management that help teams prevent issues before they happen. In the future, AI will act as a governance assistant. It will learn from past incidents, predict potential failures, and even suggest fixes in real time. This means less firefighting and more confidence in data quality. As AI continues to evolve, governance will shift from enforcing rules to enabling teams to work smarter with trustworthy, self-healing data systems.
If you’re ready to transform the way your organization handles data governance, don’t miss Kateřina’s session at the Data Innovation Summit 2025! With her experience at Kiwi.com, she’ll share insights on how to prevent data chaos and build reliable, scalable systems.
Kateřina will dive into strategies for embedding governance into engineering workflows, including data contracts, semantic layers, and AI-driven tools. Whether you’re tackling data debt or looking for proactive ways to manage your data ecosystem, her session will help you stay ahead of the curve.
Join us to learn how to shift from firefighting data issues to creating intelligent, self-healing data systems that empower your team and drive innovation forward!
Add comment