For those who’ve been anyplace close to an information crew, you already know the existential disaster taking place proper now. Listed below are just some questions information leaders and our companions have shared with us:
- Why does information governance nonetheless really feel like a slog?
- Can AI repair it, or is it making issues worse?
- How will we transfer from governance as a roadblock to governance as an enabler?
These have been the massive questions tackled on this yr’s Nice Knowledge Debate, the place a powerhouse panel of knowledge and AI leaders dove deep into dove deep into how governance must evolve.
Meet the Specialists
This dialogue introduced collectively trade leaders with deep experience in information governance, automation, and AI:
Tiankai Feng, Director of Knowledge & AI Technique at ThoughtWorks, advocates for human-centered governance and explores this philosophy in his ebook Humanizing Knowledge Technique.
Sunil Soares, founder and CEO of Your Knowledge Join, focuses on AI governance and regulatory compliance, navigating the challenges of huge language fashions in fashionable information methods.
Sonali Bhavsar, International Knowledge & Administration Lead at Accenture, drives governance methods for enterprise AI, emphasizing the significance of embedding governance from the beginning.
Bojan Ciric, Expertise Fellow at Deloitte, focuses on automating governance in extremely regulated industries, significantly monetary companies and AI-driven transformation.
Brian Ames, Head of Transformation & Enablement at Common Motors, ensures information belief as GM evolves into an AI-powered, software-driven firm.
The Three Largest Knowledge Governance Issues—And Methods to Repair Them
If there’s one factor that turned clear, it’s that governance is at a crossroads. The previous means—heavy documentation, inflexible insurance policies, and reactive fixes—merely doesn’t work in an AI-driven world. Organizations are struggling to maintain up, and governance groups are sometimes seen as roadblocks as an alternative of enablers.
However why does governance maintain failing? And extra importantly, how will we repair it? The panelists zeroed in on three main issues — and the sensible steps organizations must take to get governance proper.
1. Knowledge Governance Is All the time an Afterthought
“Governance often solely turns into vital as soon as it’s just a little too late. One thing has damaged, the info is improper, and instantly everybody realizes, ‘Oh, we must always have executed governance.’” – Tiankai Feng
Let’s be sincere: nobody cares about governance till one thing breaks. It’s the factor that will get ignored—till a foul choice, compliance failure, or AI catastrophe forces management to concentrate.
This reactive method is a dropping recreation. When governance is handled as a last-minute repair, the harm is already executed. The problem, then, is shifting governance from an afterthought to an integral a part of how organizations function.
Methods to Make Governance Proactive, Not Reactive
- Make governance an enabler, not a clean-up crew. As an alternative of reacting to issues, governance needs to be constructed into processes from the beginning. Brian Ames defined how GM reframes governance as “devour with confidence” quite than imposing top-down guidelines. The purpose? Ensuring groups can belief the info they depend on.
- Begin small and win early. As an alternative of rolling out governance throughout the whole group, give attention to a single, high-visibility challenge the place governance can ship quick worth. As Tiankai put it, “Knowledge governance takes time, however management expects immediate outcomes. It’s a must to present affect rapidly.”
- Tie governance to enterprise outcomes. If governance is barely about compliance, it would all the time be underfunded and deprioritized. Sunil Soares defined that profitable governance packages are immediately tied to income, danger discount, or price financial savings. If governance isn’t making or saving cash, nobody will care.
2. AI Is Exposing—and Amplifying—Unhealthy Governance
“AI governance is exponentially more durable than information governance. Not solely do you want good information, however now you must navigate rules, explainability, and the dangers of automation.” – Sunil Soares
The second AI entered the chat, governance bought even more durable. AI fashions don’t simply use information—they amplify its flaws. In case your information is biased, incomplete, or lacks lineage, AI will amplify these points, making unreliable choices at scale.
AI governance isn’t nearly making certain high quality information — it’s additionally about managing totally new dangers:
- Knowledge bias: AI fashions make unhealthy choices when skilled on unhealthy information. In case your information has blind spots, so will your AI.
- Lack of explainability: Many AI fashions act as “black containers,” making it not possible to grasp why they make sure predictions or suggestions.
- Automated chaos: AI brokers at the moment are making choices autonomously, typically with out human oversight. As Sunil warned, “The rules are nonetheless speaking about ‘human-in-the-loop,’ however AI brokers are actively working to take away people from the loop.”
Methods to Govern AI Earlier than It Governs You
- Take a proactive method to AI governance. Governance groups should anticipate dangers quite than scramble to repair them after an AI-driven failure. This implies aligning AI governance insurance policies with present regulatory frameworks and inside danger administration methods.
- Automate governance wherever doable. AI can truly assist repair governance by auto-documenting metadata, lineage, and insurance policies. “If governance is just too handbook, individuals received’t do it,” Bojan Ciric famous. “Automating metadata technology and anomaly detection saves time and makes governance sustainable.”
- Outline AI guardrails earlier than you want them. Organizations should create clear insurance policies outlining what AI can and may’t do. This contains monitoring AI-driven choices, implementing retention insurance policies, and making certain AI outputs are correct and explainable. Brian Ames described GM’s method: “We have to outline what our AI ‘voice’ can and can’t say. What’s its kindness metric? What are the issues it must not ever do? Governance wants to make sure AI aligns with the corporate’s model and values.”
3. No One Needs to “Do” Governance—So Make It Invisible
“For those who lead with the phrase ‘governance,’ you’re going to run into resistance. The historical past of governance is that it’s painful, bureaucratic, and irritating. We have to reframe it as one thing that allows individuals, not slows them down.” – Brian Ames
No one needs to be an information steward if it means spending half their time documenting guidelines in Excel. The most important cause governance fails? It’s too handbook, too sluggish, and too disconnected from the instruments individuals truly use.
The truth is, governance can’t depend on handbook processes. Individuals don’t need to fill out spreadsheets or sit in governance boards that really feel disconnected from their day by day work.
Methods to Construct Governance That Works, With out Anybody Noticing
- Make governance run within the background. Governance ought to occur robotically—issues like lineage monitoring, metadata assortment, and coverage enforcement needs to be constructed into workflows, not require further effort.
- Deliver governance to the place individuals already work. As an alternative of constructing groups log right into a separate governance platform, combine governance into the instruments they already use—Slack, BI platforms, engineering workflows. If governance isn’t embedded, it received’t get adopted.
- Use AI to take the burden off people. AI can generate metadata, detect anomalies, and automate compliance duties so individuals don’t must. As Sunil put it, “Individuals don’t need to do governance manually anymore—they count on AI to do it for them.”
Ultimate Takeaways: Methods to Really Make Governance Work
Governance is at a turning level. As AI reshapes how organizations use information, the previous methods—handbook, inflexible, and siloed—received’t survive. The Nice Knowledge Debate 2025 made one factor clear: governance executed proper isn’t simply vital—it’s a aggressive benefit.
The important thing to creating it work?
- Embed governance into day by day workflows. Governance can’t be a standalone course of—it should be woven into the instruments individuals already use, with automation dealing with compliance, lineage monitoring, and coverage enforcement within the background.
- Let AI govern AI. As AI adoption grows, it would tackle an even bigger position in monitoring insurance policies, detecting violations, and making certain transparency—decreasing the burden on information groups whereas stopping AI from making unchecked, high-stakes choices.
- Tie governance to measurable enterprise affect. As an alternative of being seen as a value, governance might be evaluated by its potential to guard income, enhance effectivity, and guarantee AI reliability. Organizations that show governance delivers monetary worth will acquire management assist, whereas others wrestle to safe buy-in.
- Put money into AI governance—now. Firms that delay will face mounting dangers—regulatory, reputational, and operational. As Brian Ames put it, “AI governance isn’t optionally available—it’s the inspiration for every thing we do subsequent.”
The way forward for governance isn’t nearly compliance—it’s about scaling AI responsibly and unlocking information’s full potential.
Able to construct AI-ready governance?
Atlan makes governance seamless, automated, and constructed for the AI period. Ebook a demo at present to see how Atlan will help your group scale governance with out the friction.
