Why the future of AI isn’t about better models -- it’s about better governance [Q&A]
The rise of generative and agentic AI is transforming how data is accessed and used, not just by humans but by non-human AI agents acting on their behalf. This shift is driving an unprecedented surge in data access demands, creating a governance challenge at a scale that traditional methods can’t handle.
If organizations can’t match the surge in access requests, innovation will stall, compliance risks will spike, and organizations will reach a breaking point. Joe Regensburger, VP of research at Immuta, argues that the solution isn’t more powerful AI models; it’s better governance. We talked to him to learn more.
BN: How are AI agents proliferating, and in what areas of business do you see them having the most value?
JR: AI agents are particularly valuable in expanding the visibility of resources in large, dispersed organizations. Synthesizing and cataloging vast institutional knowledge traditionally requires those associated with it to have long-term experience and immersion within an organization. AI agents have proven very effective at being able to organize and marshal these resources to increase their visibility and enhance their utility to address more business challenges, while reducing barriers. This empowers a larger cross-section of associates, particularly new associates who are just getting acclimated to the organization. Soon, AI agents may be empowered to act on behalf of human users, automating repetitive tasks, democratizing data access, improving insights, and empowering decision-making. This will enable businesses to scale operations while reducing manual effort.
BN: In what ways is the rise of generative and agentic AI transforming how data is accessed and used?
JR: The democratization of data access and utilization promises great potential. Traditionally, this has been hindered by both technical and institutional barriers. Agentic AI is proving a key enabling technology in overcoming both. On the technical end, large language models (LLMs), specifically natural language to code translation, allow users to simply ask questions, and the LLMs have proven effective at translating even complex questions into effective, machine-executable queries. But more impactful is that AI agents are proving to be effective in breaking down institutional barriers. By giving the agents a broad set of tools, like schemas, APIs, documentation, and applications, the agent can identify and leverage a broad set of often obscure tools to aid associates. The implication is that the scale of data access provisioning isn’t just increasing incrementally at a rate of 10x anymore. With AI agents, it’s skyrocketing at a rate of 100x or more, seemingly overnight. AI agents don’t just consume data -- they depend on continuous access to train, refine, and optimize their output in real time. It’s really a catch-22 scenario because AI also offers a solution to that dilemma, but I’ll get into that later in this interview.
BN: How does data access and governance need to evolve to keep up?
JR: Traditional governance models assume human control over data access, but AI agents require a more adaptive approach. Soon, governance will need to account for both human and non-human, AI-driven access decisions, with policies that adjust dynamically based on roles, responsibilities, and real-time business context. This means shifting to policy-driven automation that scales governance at machine speed. Rather than replacing governance teams, this approach eliminates manual and tedious approvals, freeing teams to focus on strategy and oversight. AI is accelerating the volume, velocity, and complexity of data interactions, completely doing away with traditional governance models. Attempting to keep up isn’t enough -- data governance must reach a place where it’s automated, intelligent, and built for scale.
BN: The rise of AI agents is clearly going to put a strain on data governors – how can organizations use AI in data access and governance to keep up with this demand?
JR: The rise of AI agents is accelerating data access demands at a scale that governance teams were never built to handle. Requests are exploding, moving beyond what human-driven oversight can manage. In the case of data governance and provisioning, AI can help to streamline data access request approvals and become progressively advanced by recommending policies and automating decision-making. This means businesses can move faster, discover data more easily, and streamline processes that require cross-functional collaboration. This is just the beginning -- as governance scales, AI will play an even greater role in automating enforcement, monitoring access at scale, and ensuring security and compliance in real-time. But as organizations race to harness their potential, it’s important that they balance speed with control, ensuring flexibility while maintaining fail-safe protections. Without governance that can scale with AI, enterprises risk losing control, not just over access, but over how data is used, shared, and acted upon. The future of AI isn’t just about better models -- it’s about better data governance.
BN: What data governance tasks will remain with humans, or will no tasks remain exclusively with humans?
JR: While AI agents will help with many operational governance tasks, humans will still play a crucial role in setting strategy, defining ethical guidelines, and making high-stakes decisions. AI is an augmentation tool, not a full replacement for humans. Human oversight will remain essential for ensuring that AI-driven actions align with business and regulatory objectives.
Image credit: Tongsupatman/Dreamstime.com