How to innovate responsibly in AI: Insights from Bank of America and Deloitte

GenAI won’t succeed on models alone. Without strategic governance — and the ability to provision governed access at scale — organizations risk bias, errors, and compliance failures. This Forbes piece underscores why treating data as an afterthought is no longer an option. The winners in AI will be the ones who make data provisioning the strategy. 🔗 https://ow.ly/8ssP50XlHlX #DataProvisioning #GenAI #AIInnovation #DataStrategy

I saw a sticker once that said, "Seize the data!" True that, on a few levels. Here's why. So, here is the food for thought....and I've been saying this for a while to drive the points home. If 79% of organizations view AI as vital to their future business success, or dare I say business viability, why is that 14% do not have the data maturity to fully take advantage of their data? Things like data provenance, location, context, governance, are often given short shrift by organizations only to find out that the cost of fixing these issues can later cost them exponentially. This is why most implementations fail (80%+ according to Gartner and others), and the cost of the failures averages $15M per (again, according to Gartner and others). The cost of a good solution for this is far less than the $15M! While there may be others, I think solutions like #Cyera can get an organization on the right path quickly, from the start. https://lnkd.in/d6eMHNKT

Some key notes - 1. 30% of GenAI POCs will be abandoned by end 2025. 2. GenAI becomes a powerful tool when organisations transform chaotic, unstructured data into a structured, governed knowledge base. 3. Avoid falling into the trap of “AI for AI’s sake.” https://lnkd.in/grhR8a3J

How One Chief Data & AI Officer Ships 'Governance-as-a-Product' to Prove ROI Many enterprises still treat data governance as a cost center that slows work while AI and compliance requirements gain speed. Jeremy Stierwalt, CDAIO at Paradigm Technology, explains how taking a "governance-as-a-product" ap…

Title: Celebrating Data Integrity Champions: How 2025 Awardees Are Shaping AI for a Better Tomorrow Content: In an era where data drives innovation, Precisely's announcement of the 2025 Data Integrity Award winners highlights a pivotal shift toward harnessing trustworthy information to unlock meaningful AI advancements. These organizations exemplify how maintaining data accuracy and reliability is not just a technical necessity but a foundational driver for impactful business strategies and societal progress. By shining a spotlight on those who excel in data integrity, Precisely underscores the critical role that clean, verifiable data plays in fostering AI solutions that can be confidently applied in real-world scenarios. Among the winners, a remarkable thread emerges: the dedication to using AI not only for commercial gain but also to tackle some of the most pressing global challenges. From climate risk assessment to sustainability reporting, these innovators are leveraging high-integrity data sets to generate insights that empower decision-makers to act responsibly and proactively. This approach fosters transparency and accountability, setting a strong example for industries aiming to align their operations with environmental and social governance principles. What stands out is the commitment to inclusivity evidenced by several awardees who utilize trusted data to bridge gaps for underserved populations. Access to vital services, whether healthcare, finance, or education, often depends on the quality and reliability of underlying data. These organizations are demonstrating that data integrity is foundational to creating equitable systems where technology serves as a democratizing force. By ensuring accurate data flows into AI algorithms, they enable better outcomes and open new avenues for social upliftment. From a strategic lens, the recognition of these cutting-edge applications reflects a growing awareness that data governance and ethical stewardship are integral to the future of AI innovation. As businesses navigate increasing scrutiny from regulators, customers, and communities, the emphasis on data integrity provides a competitive edge and builds lasting trust. It also signals a broader industry evolution where success is measured not simply by technological breakthroughs but by demonstrable positive impact and societal benefit. Looking forward, Precisely's 2025 award honorees serve as inspiring examples of how data-driven innovation can harmonize profit with purpose. Their achievements make clear that investing in robust data management practices is not a peripheral concern but a core strategic imperative. In a world increasingly reliant on AI, these leaders prove that integrity in data is the pathway to unlocking trustworthy intelligence that drives sustainable growth and meaningful change.

Data governance is not just about policies and technology — it’s about people, trust, and dialogue. In many organisations, informal power dynamics and competing interests shape how data is managed and shared. Addressing this requires more than frameworks; it needs a culture of transparency and shared ownership. As studies highlight (Data4SDGs, 2021; TDWI, 2023), sustainable governance succeeds when institutions engage all stakeholders and align around clear values. When we treat governance as collaboration — not control — data finally serves its true purpose: empowering informed, ethical, and inclusive decisions. Sources: Data4SDGs. A roadmap for putting values at the heart of data (2021) – [data4sdgs.org https://lnkd.in/dxvMcFKB TDWI. Ten Principles of Effective Data Governance* (2023) – tdwi.org https://lnkd.in/duwqfnzs

📢Check out this blog from IDinsight head of data engineering Eric Dodge: "To unlock the value of AI, you need data infrastructure." We often say that AI tools are "garbage in, garbage out" without good, interoperable, clean, and up-to-date data systems undergirding them. This article goes deeper into the technical underpinnings of all good AI tools, and issues a call to action for doers and funders alike to invest in healthy data systems. As Eric points out, the real challenge isn’t just technical. Organizational change management and leadership are just as critical to realizing the promise of new technologies. https://lnkd.in/eE98k_E2

What Data Governance Really Means in 2025 Most companies think data governance is about control. It is not. It is about trust. Governance isn’t a roadblock. It is the bridge between chaos and confidence. In 2025, true governance means: Clear roles and ownership Traceable data lineage Documented metadata A culture of stewardship Without these, data strategy becomes noise, dashboards lose meaning, and compliance becomes a nightmare. Don’t wait for regulators. Start small, but start now. #DataGovernance #DataTrust #DataStrategy #Metadata #AI #DigitalTransformation

💡 Could #Data Be Fragile? Probably not. But the way we organize data - what I call data #anthology - definitely can be. Many of you know Nassim Nicholas Taleb, the Lebanese philosopher who developed the concept of fragile 💎 – resilient 🛡– antifragile ⛓ systems. His framework describes how some structures break under stress, others resist it, and a few actually benefit from it. Over the past few weeks, I’ve had several conversations about data. One of them was a debate on data management in a world defined by #change. And I’ve become convinced that Taleb’s antifragility applies directly to data anthologies. 🧩 From Fragile to Antifragile Data Management Traditional data management sits on a spectrum between chaos and rigid order. 📍On one side, minimal taxonomy leads to disorganization - data is inconsistent, scattered, and valuable insights are lost. 📍On the other, hyper-controlled taxonomies demand endless manual effort, discipline, and time. Both are fragile because they collapse under the pressure of change. They can’t evolve fast enough. But in an antifragile model, the system would embraces change and learn from it. Every semantic conflict, user correction, or classification mismatch becomes a signal that improves the system. Is this possible? 🤖 AI as the Catalyst Yes – and it is #AI what makes this vision realistic. Of course not as a replacement for well-designed processes or human judgment, but as an augmenter, a dynamic layer that continuously evolves data organization. Imagine an AI-powered system that: 📍Regularly suggests taxonomy updates to data stewards 📍Accepts user feedback and tests temporary rules 📍Promotes successful adaptations into standards 📍Monitors the impact of every change on accuracy and quality This would transform static, “dictionary-style” data catalogs into living, federated frameworks - decentralized, adaptive, and self-improving. 🌱 The Future of Data Anthology To make data antifragile, we need to reimagine taxonomy itself - from rigid hierarchies to multi-layered, evolving ecosystems. Systems that not only survive change but grow stronger because of it. That’s the real promise of AI in data management. Not automation for its own sake, but collaboration between human expertise and adaptive intelligence - a partnership that learns, evolves, and thrives in complexity. Of course this is a thought exercise. #knowledge #knowledgemanagement #datamanagement #transformation #PwC

📬 This post is part of the “Data Management in a Nutshell: Insight Series”—short, high-impact takeaways from the full Data Management in a Nutshell newsletter, now followed by over 6,400 professionals worldwide. 📩 Subscribe to get the full version delivered to your inbox: https://lnkd.in/dmZHyAm2 🎥 Watch the full presentation: https://lnkd.in/e7RM6Ez8 🔎 Why harmonizing data and AI governance matters Many organizations today manage data and AI through separate frameworks—often developed by different teams, driven by distinct regulations, and guided by unrelated priorities. This fragmentation leads to duplicated efforts, inconsistent controls, and a lack of shared understanding about what’s really being governed. Yet, both frameworks rely on the same foundation: data. The question is no longer whether they should be connected, but how to align them effectively without creating unnecessary complexity. 📊 What influences the level of alignment Four key factors shape this decision. The first is the definition of data-related concepts: what counts as data, metadata, or information, and whether AI-generated data falls under the same governance scope. The second is AI-related concepts: how we define AI systems, models, and outputs, and how they relate to data products and assets. The third and fourth factors are regulations and frameworks—both data-related and AI-related—which differ widely across jurisdictions and industries. Each organization must interpret these factors and decide how closely its data and AI governance should operate. 🧭 How organizations can make the right decision The most effective approach starts with scoping—defining which data and AI activities fall within the framework. Next, organizations need a structured way to analyze dependencies between capabilities, design integrated governance structures, and implement policies and processes that are adaptable and scalable. Finally, performance and maturity measurement ensure that governance continues to evolve as the business and technology landscape change. When applied consistently, this approach turns alignment from a theoretical goal into a practical, measurable capability. 💬 In short Harmonization isn’t about merging frameworks—it’s about making them work together, guided by common principles, a shared understanding of data, and coordinated execution. Join 🗓️ New Upcoming Event: Q&A Live Session: Data Lineage in Practice https://lnkd.in/e_em3Ecs #data #AI #governance #framework