Toon Weyens, Ph.D.’s Post

I once bragged about a database with 150 hand-crafted edit checks. In 2026, an AI-driven system would generate most of those automatically. That insight is at the heart of a new blog post from Herb Blecher at Enterprise Management Associates (EMA) on building a modern data and analytics practice in 2026. Data quality has evolved from manual rules to observability and now to AI-driven data reliability. The real question is no longer “Did it pass my rules?” but “Is the data behaving normally, and how fast can we fix it?” Herb explores how quality, observability, governance, lineage, security, and cost are rapidly converging and what that means for data teams this year. Read the full blog: https://lnkd.in/gN3CehMZ If you’re a vendor or practitioner navigating this shift, we’d love to hear what you’re seeing. What’s working? What’s falling short? What problems still don’t have real answers?

Introducing DataAgent, free tools for data geeks In today’s data-driven world, organisations are generating and consuming more data than ever before. With this explosion of data comes a critical challenge: how do you maintain quality, trace lineage, and ensure governance across complex data ecosystems? This is where TraceTrail comes in – a comprehensive data quality and lineage tracking platform designed to help data teams identify, document, track, and resolve data quality issues across complex data ecosystems. Read more about it here: https://lnkd.in/deQ7N2ij We are excited to hear your thoughts, so please feel free to check out the repo for instructions on how to run it locally and experience the app with demo data. https://lnkd.in/dQj43Rmz

🎄 5 must-reads on Data Governance for Christmas break : 1️⃣ The decline of data organizations https://lnkd.in/e9tFDYMH 2️⃣ Difference between semantics, ontology and taxonomy https://lnkd.in/eP7fykgn 3️⃣ Who owns data quality anyway? https://lnkd.in/ebPJ7kKR 4️⃣ AI-ready data : a technical assessment https://lnkd.in/drkiVBnJ 5️⃣ Data catalogs 2.0 : get value from metadata https://lnkd.in/eicEkHHD Merry Christmas !

Great 2026 predictions blog by Principal Analyst, Stephen Catanzano. "2026 will not be defined by a single breakthrough or architectural change. It will be defined by a shift in mindset. Data management stops being treated as supporting infrastructure and starts being recognized as the system that enables intelligence itself." https://lnkd.in/eGvJSifD

You guys ever read something and wonder "did he just steal my notepad?" Kinda felt that way when I came across my buddy Stephen Catanzano's latest opinion article on what we all have to look forward to in 2026. Three big things that particularly hit home for me? 1. Consolidated, unified platforms aren't the desination... they're a starting point. Enterprises are trying to reduce vendor sprawl and keep up with the speed and scale AI requires. The key is the level of integration of your solutions. Depth of integration leads directly to being able to keep up (and duplicate the success). 2. Data Products are growing up. They're not just curated data sets anymore. They're operational building block. The momentum they experienced in 2025 is going to grow... and the benefit from ensuring context, governance, trust and reusability is NOT to be underestimated. 3. You partners matter now more than ever. The market is moving too fast for enterprises to bet big on someone that they don't trust to (a)meet them where they are and (b) have a clear vision of how to get them to where they need to be. If you're in the data management space, this one is definitely worth the read. And stay tuned for more soon. https://okt.to/S4edwu

Just published in Forbes: "Why Data Quality Should Start at the Beginning, Not the End." Organizations often treat data quality as an afterthought, addressing it only at the end of the data life cycle. However, retrofitting quality into bad data is akin to trying to unbake a cake—it's costly, time-consuming, and rarely effective. The true return on investment lies in establishing quality from the outset: 1. Data Testing during development 2. Real-time Monitoring with Data Controls in production pipelines 3. Data Quality and Observability on the final data Read the article for key takeaways: https://lnkd.in/edrhu347 Do you agree with my thesis? What percentage of your data budget is allocated to measuring and cleanup versus prevention? Share your experience below! #iceDQ #DataQuality #DataGovernance #DataStrategy #Forbes #DataEngineering #DataReliability

Approaching governance for unstructured data versus structured data requires new skills and tools that are now maturing. By Komprise COO Krishna Subramanian

At first, there is only noise. Unlabeled data streams in from everywhere. Logs, signals, text, numbers. It has no intent. No memory. No direction. On its own, it does nothing. Next, "structure appears and logic begins to shape the noise. Rules form. Models infer. Decisions become possible, not because the system is intelligent yet, but because it has learned how to reason. Logic does not think. It evaluates. This issue of The Pulse follows that same path. From data to logic. From memory to behavior. From intelligence to governance. Not as isolated topics, but as a single system that either works together or fails together. Artificial intelligence is not a feature. It is a story of constraints, choices, and execution. It doesn’t begin with intelligence. That’s the mistake most systems make. They chase capability before coherence, scale before control. They stack models on top of data and call it progress, only to discover later that no one knows why it behaves the way it does. This system takes a different path. This is where intelligence is allowed to emerge only after the foundations are set. Data is questioned before it is trusted. Logic is tested before it is deployed. Memory is scoped, not hoarded. Behavior is observed, not assumed, and every step leaves a trace. Why say you? Because intelligence without traceability is indistinguishable from error. As the system grows, pressure builds. More data arrives. More decisions are possible. Optimization begins to pull in directions no single engineer explicitly chose. This is the moment where most architectures fracture. But governance is already there. Not as a document. Not as a committee. As code and constraints live beside logic. Boundaries are enforced at execution time, and not reviewed after the fact. When the system acts, it does so within rules that are as real as the functions that trigger the action. Data doesn’t act on its own. Memory doesn’t persist accidentally. Every transition from intent to execution passes through code that can be read, audited, versioned, and changed. That bridge matters because when something breaks, the question is never “what did the model do?” The question is “why was it allowed to do it?” In this system, the answer exists. The Pulse follows that thread deliberately. Each section returns to the same question from a different angle: How do we design intelligence that remains accountable as it scales? Not smarter systems. Not faster ones. Systems that can explain themselves. This is not a manifesto. It’s a field report. It's a record of what happens when artificial intelligence is treated as an engineered system instead of a spectacle. This story doesn’t end here. It stabilizes, and from that stability, intelligence finally becomes useful. The Pulse isn’t finished when it’s published. It advances through dialogue, scrutiny, and most importantly, real-world use. That’s the call. Want to learn more? Join the conversation.

At first, there is only noise. Unlabeled data streams in from everywhere. Logs, signals, text, numbers. It has no intent. No memory. No direction. On its own, it does nothing. Next, "structure appears and logic begins to shape the noise. Rules form. Models infer. Decisions become possible, not because the system is intelligent yet, but because it has learned how to reason. Logic does not think. It evaluates. This issue of The Pulse follows that same path. From data to logic. From memory to behavior. From intelligence to governance. Not as isolated topics, but as a single system that either works together or fails together. Artificial intelligence is not a feature. It is a story of constraints, choices, and execution. It doesn’t begin with intelligence. That’s the mistake most systems make. They chase capability before coherence, scale before control. They stack models on top of data and call it progress, only to discover later that no one knows why it behaves the way it does. This system takes a different path. This is where intelligence is allowed to emerge only after the foundations are set. Data is questioned before it is trusted. Logic is tested before it is deployed. Memory is scoped, not hoarded. Behavior is observed, not assumed, and every step leaves a trace. Why say you? Because intelligence without traceability is indistinguishable from error. As the system grows, pressure builds. More data arrives. More decisions are possible. Optimization begins to pull in directions no single engineer explicitly chose. This is the moment where most architectures fracture. But governance is already there. Not as a document. Not as a committee. As code and constraints live beside logic. Boundaries are enforced at execution time, and not reviewed after the fact. When the system acts, it does so within rules that are as real as the functions that trigger the action. Data doesn’t act on its own. Memory doesn’t persist accidentally. Every transition from intent to execution passes through code that can be read, audited, versioned, and changed. That bridge matters because when something breaks, the question is never “what did the model do?” The question is “why was it allowed to do it?” In this system, the answer exists. The Pulse follows that thread deliberately. Each section returns to the same question from a different angle: How do we design intelligence that remains accountable as it scales? Not smarter systems. Not faster ones. Systems that can explain themselves. This is not a manifesto. It’s a field report. It's a record of what happens when artificial intelligence is treated as an engineered system instead of a spectacle. This story doesn’t end here. It stabilizes, and from that stability, intelligence finally becomes useful. The Pulse isn’t finished when it’s published. It advances through dialogue, scrutiny, and most importantly, real-world use. That’s the call. Want to learn more? Join the conversation.