Aligning Info Management & Data Management for AI Adoption

Really looking forward to this one.   Su Jella MAICD, MBA and I are teaming up with Adelle Ford and Annette Senior from Recordkeeping Innovation to talk about how information management and data management are converging.   We'll be discussing metadata, governance, lineage, risk across structured and unstructured data.   I've seen organisations spend heavily on AI only to realise they never connected their records management to their data governance. It gets messy, fast. Don't let that be you.   Set the right foundation. Join us tomorrow from 11AM - 12PM AEST.   Register here: https://lnkd.in/g-HBsVMk Use code: IMSFREE when you register.

IM Systems is delighted to be partnering with Recordkeeping Innovation to present an educational webinar for RIMPA Global next Tuesday, 19 May 2026 from 11am–12pm AEST. “Aligning Information Management and Data Management Practices” As organisations accelerate AI adoption, the alignment between Information Management, Records Management, Data Governance, and AI Governance is becoming increasingly important. The webinar will be presented and facilitated by George Bassili and Su Jella from IM Systems, together with Adelle Ford and Annette Senior from Recordkeeping Innovation. Together, we will explore: • How Information Management and Data Management are converging • The importance of metadata, governance, and lineage • Managing information risk across structured and unstructured data • Building trusted foundations for analytics and AI The session will also include live audience interaction using Mentimeter, along with discussion around the practical challenges organisations are facing as they modernise their governance and AI capabilities. We’re also thankful to RIMPA for providing a complimentary registration code for our clients, colleagues, and friends. Use code: IMSFREE Register here: https://lnkd.in/gayb8FK2 We look forward to joining Recordkeeping Innovation and the broader RIMPA community for the discussion. #RIMPA #InformationManagement #RecordsManagement #DataGovernance #AI #AIGovernance #Metadata #InformationGovernance

AI will not fix poor data governance. It will expose it. In many public sector environments, data governance is still treated as a secondary task assigned to officials who already have full operational responsibilities. They are expected to support reporting, correct data quality issues, respond to audit queries, assist system projects, clarify definitions, and resolve inconsistencies — often without formal authority, dedicated capacity, or clear institutional ownership. That model is no longer sustainable. As government departments modernise systems, integrate platforms, improve reporting, and begin exploring AI-enabled services, the quality and governance of data becomes a delivery issue — not only a compliance issue. For public service departments, the DPSA’s Determination and Directive on the Implementation of Data Governance in the Public Service, supported by the Public Service Data Governance Framework, reinforces this shift. Data governance requires more than goodwill. It requires: • Defined roles • Clear ownership • Maturity assessment • Executive oversight • Enabled data stewards • Business areas that understand their accountability Without this, institutions risk building modern platforms on weak information foundations. The result? • Reports that cannot be trusted • Systems that do not align • AI outputs that cannot be explained • Audit evidence that is difficult to defend • Decisions based on inconsistent information Digital transformation depends on trusted data. AI readiness depends on governed data. Public value depends on information that is accurate, owned, protected, and usable. The real question is no longer whether data governance is important. The question is whether it is still being treated as an additional task — or as a core institutional capability. Where do you see the biggest data governance gap in government: ownership, quality, roles, maturity, accountability, or executive oversight? #DataGovernance #AIReadiness #DigitalTransformation #PublicSector #ICTGovernance #DigitalGovernment #GovTech

𝐘𝐨𝐮 𝐜𝐚𝐧 𝐡𝐚𝐯𝐞 𝐭𝐡𝐞 𝐛𝐞𝐬𝐭 𝐩𝐞𝐨𝐩𝐥𝐞 𝐚𝐧𝐝 𝐠𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞 𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞𝐬 𝐢𝐧 𝐩𝐥𝐚𝐜𝐞. 𝐁𝐮𝐭 𝐢𝐟 𝐭𝐡𝐞 𝐝𝐚𝐭𝐚 𝐢𝐭𝐬𝐞𝐥𝐟 𝐜𝐚𝐧𝐧𝐨𝐭 𝐛𝐞 𝐭𝐫𝐮𝐬𝐭𝐞𝐝, 𝐧𝐨𝐧𝐞 𝐨𝐟 𝐢𝐭 𝐦𝐚𝐭𝐭𝐞𝐫𝐬. In Post 2 we explored the human side of Data Governance — ownership, stewardship, and the structures that make accountability work. Today we turn to the data itself — the practices and processes that make data trustworthy, discoverable, and fit for purpose. Because governance is only as strong as the data it protects. 📍 𝐃𝐚𝐭𝐚 𝐐𝐮𝐚𝐥𝐢𝐭𝐲 → Accurate, complete, consistent, and timely — these four dimensions define whether data is fit for use → Data quality is not a one-time fix — it requires continuous monitoring, clear standards, and ownership at the source 📍 𝐌𝐞𝐭𝐚𝐝𝐚𝐭𝐚 & 𝐃𝐚𝐭𝐚 𝐂𝐚𝐭𝐚𝐥𝐨𝐠𝐮𝐢𝐧𝐠 → Metadata is data about your data — it tells you what a dataset contains, where it came from, and how it should be used → A well-maintained data catalogue makes data discoverable across the organization, reducing duplication and building trust 📍 𝐃𝐚𝐭𝐚 𝐋𝐢𝐧𝐞𝐚𝐠𝐞 → Lineage tracks the journey of data — from its origin, through every transformation, to its final destination → Understanding lineage is essential for troubleshooting quality issues, meeting regulatory requirements, and governing AI models responsibly 📍 𝐌𝐚𝐬𝐭𝐞𝐫 𝐃𝐚𝐭𝐚 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 → Master data — customers, products, locations, suppliers — is the core reference data the whole organization depends on → Managing it consistently across systems ensures everyone is working from a single, trusted version of the truth 📍 𝐖𝐡𝐲 𝐭𝐡𝐢𝐬 𝐦𝐚𝐭𝐭𝐞𝐫𝐬 𝐟𝐨𝐫 𝐀𝐈  → AI models are only as reliable as the data pipelines feeding them — lineage, quality, and cataloguing are prerequisites, not optional extras → Organizations that cannot trace and trust their data will struggle to explain, audit, or improve their AI outputs Data practices done well do not just support governance — they become the backbone of every analytical and AI initiative the organization undertakes. 🔵 ❖ What has been the biggest data quality or metadata challenge you have faced — and what made the difference in addressing it? #DataGovernance #DataQuality #DataLineage #MasterDataManagement #AIReadiness

Most organisations do not have a data problem. They have a data management problem. The challenge is rarely a lack of information. It is fragmented systems, inconsistent ownership, unclear governance, and data that cannot be trusted when decisions matter. Our Data Management Canvas was designed to help organisations step back and assess the foundations properly: • Data quality • Governance • Structure and interoperability • Ownership and accountability • AI readiness It is a practical framework for teams trying to move from disconnected data activity to a coherent strategy. If AI is part of your roadmap, data management cannot be an afterthought. Reliable AI depends on reliable data. Explore the Data Management Canvas: https://lnkd.in/e9bBaZZ5

One thing I’ve observed across enterprise data environments is that many “data quality” problems are actually awareness, ownership, and communication problems expressed through data. In many organisations, the issue is not simply that: • a field is null • a transformation failed • a report contains inconsistent numbers The deeper issue is often that: • business owners lack visibility into downstream impact • engineering teams lack operational business context • ownership boundaries are unclear • governance exists separately from operational delivery In many environments, data quality discussions become heavily tooling-focused: • rules • dashboards • alerts • scorecards • monitoring platforms Those things matter — and I’ve led many of those conversations myself. But tooling alone rarely resolves trust issues when the surrounding operational model remains fragmented. The more effective data environments I’ve worked around tend to share a few characteristics: • clear ownership and accountability • shared business and technical understanding • visibility into downstream impact • governance embedded within engineering and delivery processes • feedback loops connecting operational issues back to business outcomes What’s interesting is that AI and advanced analytics are accelerating visibility into many of these existing gaps. AI systems expose ambiguity, inconsistency, and weak operational alignment far faster than traditional reporting environments ever did. In reality, many of these issues already existed: • inconsistent definitions • fragmented ownership • unclear lineage • disconnected governance • unresolved operational dependencies AI simply makes them harder to ignore. As organisations push toward AI-enabled decision-making, unresolved ownership and semantic inconsistencies surface much faster because AI amplifies underlying uncertainty rather than quietly masking it. A governance issue can quickly become: • an operational issue • a trust issue • or ultimately a business decision-making issue Over time, I’ve become less convinced that enterprise data governance is primarily about control. And more convinced that it is about enabling operational clarity, trust, and aligned decision-making across increasingly complex systems.

Most companies still talk about Data Quality as if it is one single problem. It is not. Data Quality is actually a combination of dimensions and most organizations fail because they focus on only one of them. Usually accuracy. But data can be accurate and still completely useless for the business. I have seen customer records that were technically accurate but duplicated five times across systems. I have seen dashboards with perfect calculations built on outdated data from three days ago. I have seen financial reports where every number matched perfectly but nobody could explain what the metrics actually meant. That is why Data Quality dimensions matter. Accuracy is only one piece of the story. Completeness matters because missing values create operational blind spots. Consistency matters because different systems should not tell different versions of the truth. Timeliness matters because late data in modern business is often equivalent to bad data. Validity matters because data must follow expected formats and business rules. Uniqueness matters because duplicate records destroy trust very quickly. And relevance is becoming increasingly important because companies are collecting enormous volumes of data that nobody actually uses. One interesting pattern I keep seeing in enterprises is this: Data teams spend months building validation rules for accuracy while business users continue complaining about inconsistency and lack of context. The technical definition of quality and the business definition of quality are often completely different things. Another uncomfortable observation: Many organizations proudly report 95 percent data quality scores while their business teams still maintain parallel Excel files because they do not trust enterprise data. That number alone should make every governance leader uncomfortable. The reality is that bad data is rarely caused by technology alone. It usually comes from unclear processes, fragmented ownership, disconnected systems and constant changes in business definitions. And now with AI entering every discussion, weak Data Quality dimensions become even more dangerous. AI does not understand business context. It simply amplifies whatever quality level already exists. If the foundation is inconsistent, incomplete or outdated, the output will scale those problems faster than ever before. Data Quality is no longer a reporting issue. It is becoming a business credibility issue. I write about data, governance, and how things actually break in companies. ➤ Follow Demin Alexey if that’s your space. ♻️ Repost to help another data leader speak the language of business. #DataGovernance #Data #AI #DataStrategy #DigitalTransformation

𝐌𝐨𝐬𝐭 𝐂𝐃𝐎𝐬 𝐚𝐫𝐞 𝐛𝐞𝐢𝐧𝐠 𝐚𝐬𝐤𝐞𝐝 𝐭𝐨 𝐬𝐮𝐩𝐩𝐨𝐫𝐭 𝐭𝐡𝐞 𝐀𝐈 𝐬𝐭𝐫𝐚𝐭𝐞𝐠𝐲. 𝐒𝐨𝐦𝐞 𝐨𝐟 𝐭𝐡𝐞𝐦 𝐬𝐡𝐨𝐮𝐥𝐝 𝐛𝐞 𝐥𝐞𝐚𝐝𝐢𝐧𝐠 𝐢𝐭. One pattern I keep noticing across enterprise AI programs is that organizations often view AI as the strategic initiative and data as the dependency. Yet the foundation every agent relies on is not an AI asset. It is a data asset. → Agents do not simply retrieve information. They resolve ambiguity, infer context, and make decisions at runtime. That makes the semantic layer far more than a reporting convenience. It becomes a runtime dependency for every agent operating across the enterprise. → When definitions are inconsistent, lineage is unclear, or governance is fragmented, agents inherit that uncertainty. What I have observed in several large organizations is that delivery teams rarely wait for those issues to be resolved. → Under pressure to deliver business outcomes, they often build their own context layer, business definitions, metadata mappings, and governance controls around the use case at hand. According to Gartner, 80% of enterprise applications now embed at least one AI agent, up from 33% just two years ago. At the same time, more than 40% of Agentic AI projects are projected to be canceled by 2027, with governance, semantic inconsistency, and data readiness among the most common challenges. What I find particularly interesting is not the failure rate itself. It is what happens organizationally when the enterprise data foundation is not agent-ready. The center of gravity begins to shift. Some of the most effective data leaders I speak with are not focusing exclusively on data management. They are increasingly treating semantic context, metadata, governance, and business meaning as strategic assets that AI systems depend on. One useful thought exercise might be: 🔹 Look at your highest-priority agent initiatives 🔹 Identify the semantic definitions those agents rely on 🔹 Ask whether those definitions are governed, versioned, and owned as enterprise assets The answer often reveals how prepared the organization really is for Agentic AI. 𝐈𝐧 𝐚𝐧 𝐚𝐠𝐞𝐧𝐭-𝐟𝐢𝐫𝐬𝐭 𝐞𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞, 𝐭𝐡𝐞 𝐭𝐞𝐚𝐦 𝐭𝐡𝐚𝐭 𝐨𝐰𝐧𝐬 𝐜𝐨𝐧𝐭𝐞𝐱𝐭 𝐰𝐢𝐥𝐥 𝐨𝐟𝐭𝐞𝐧 𝐬𝐡𝐚𝐩𝐞 𝐭𝐡𝐞 𝐨𝐮𝐭𝐜𝐨𝐦𝐞𝐬. The question is whether that context is being defined intentionally today, or recreated project by project. #EnterpriseAI #DataStrategy #DataGovernance #AITransformation

Government data is also a trust, governance, and organizational alignment challenge. In the final part of this three-part CDO Magazine interview series, Karthik Yajurvedi, Chief Data Officer for the Commonwealth of Massachusetts, joins Adita Karkera, Ph.D., Chief Data Officer for Deloitte's Government & Public Services, to discuss: • Why silos remain the biggest barrier to collaboration • How Massachusetts approaches governance without creating bottlenecks • Why data quality must be evaluated in context, especially in the age of AI • What aspiring CDOs need to understand about stakeholder management and business strategy ➡️ Read the full interview: https://hubs.ly/Q04gYn3s0 #DataGovernance #AI #CDO #GovernmentData #DataLeadership #ArtificialIntelligence #PublicSector #Deloitte

Glad this idea is resonating. The shift from data governance to decision governance is something I've been thinking about for a while and what I'll be exploring further at the Dow Jones Risk Journal Briefing in Stockholm on May 28th. The question isn't just whether your data is good. It's whether your organisation has decided who decides and at what threshold a human needs to be in the loop. That's a governance problem. And most organisations haven't started solving it yet.