I joined the webinar organized by DAMA ESPAÑA and Gaia-X Hub España expecting a largely technical deep dive into data governance and data spaces. What I did not expect was how clearly it crystallized something I have seen repeatedly in real organizations: governance only feels slow when shared meaning is missing.
Too often, governance is framed as a bottleneck, the dreaded “office of NO” that delays initiatives and complicates progress. Yet throughout the session, and particularly in the way practitioners like Daniel Torbellino Rubio articulated their experience, that narrative was quietly dismantled. Governance was presented not as a brake, but as an enabler, the mechanism that allows organizations to move faster without collapsing under their own scale.
Anyone who has worked with data long enough recognizes the pattern. Early data initiatives feel agile and liberating. Definitions are implicit, datasets are spun up quickly, experimentation thrives. But as soon as scale enters the picture, fragility appears. KPIs start to mean different things depending on the team. Data products multiply without clear ownership. Compliance arrives late and under pressure. AI pilots that looked promising in demos struggle to survive contact with production.
One phrase from the webinar captured this dynamic with uncomfortable precision: “garbage in, garbage share.” Data spaces are exceptionally good at distributing data, but they are agnostic to its quality, meaning, or coherence. Without solid internal governance, they do not correct disorder; they amplify it. Connectors and catalogs simply become more efficient ways of spreading confusion.
This is where governance reveals its real value. Not as control for its own sake, but as friction reduction. When definitions are agreed upfront, discussions shorten. When quality expectations are explicit and fit for purpose, rework decreases. When responsibilities are clear, decisions accelerate. Governance reduces organizational latency, the hidden tax paid every time teams have to renegotiate meaning from scratch.
What struck me most during the webinar was how central semantics was to all of this, even when it was not always named directly. It surfaced repeatedly through references to glossaries, metadata management, shared terms, and practical annexes designed to make complexity digestible. In collaborative environments, semantics is not documentation overhead; it is operational infrastructure.
This perspective also shaped the way interoperability was discussed. Rather than advocating for forced standardization, the conversation leaned toward a more realistic and mature approach: interoperability by contract, not by homogenization. Each organization retains its internal models and language, but meaning is explicitly negotiated at the boundaries. In federated data spaces, diversity is inevitable, and governance exists to make that diversity workable rather than to erase it.
A metaphor used during the session made this tangible. Data is what is inside the package. Semantics, expressed through metadata, is the label that tells you where it comes from, how it should be handled, and whether it can be trusted. Without the label, logistics collapse. In data spaces, meaning is what enables discovery, reuse, and confidence.
Although artificial intelligence was not always explicitly foregrounded, it was present throughout the discussion. And that makes sense. AI systems rarely fail because of missing algorithms; they fail because meaning is unstable. Inconsistent categories, shifting definitions, and unclear context lead to outputs that may be statistically impressive but operationally unreliable.
Governance grounded in semantics is what anchors AI to reality. It defines what concepts actually mean, what quality is sufficient for a given purpose, who owns each data product, and which uses are legitimate. It enables auditability and accountability over time. Without this foundation, AI is little more than advanced pattern recognition without a compass.
By the end of the webinar, one idea stood out clearly. Mature governance should not be measured by how many initiatives it blocks, but by how many it enables, safely, responsibly, and at speed. Data spaces are, at their core, trust architectures. And governance is what makes that trust scalable.
I would like to close by acknowledging the work behind this webinar and the documents it builds upon. The clarity and depth of the discussion are the result of sustained, collective effort. My congratulations to Sofia Kosenko , Daniel Torbellino Rubio , Mélida López and Gemma Anglada Caballé for the rigor, generosity, and commitment reflected in this work. Contributions like these are what move data governance forward from theory into something that can genuinely enable collaboration, trust, and value at scale.