Dominic Jainy sits at the intersection of emerging technology and corporate pragmatism. As an expert in artificial intelligence and machine learning, he has spent years observing how the initial hype of digital transformation often hits a wall when it encounters the messy, fragmented reality of enterprise data. In this conversation, we explore the strategic implications of a massive shift in the industry: the move from viewing data as a mere storage challenge to treating it as the structural foundation of the business. We delve into how the acquisition of Informatica by Salesforce signalizes a new era where “trusted context” is the most valuable asset an organization can own, surpassing the models themselves in long-term importance.
The following discussion summarizes the critical evolution of Master Data Management in the age of AI agents, the competitive maneuvers of tech giants like Microsoft and Snowflake, and the practical hurdles large-scale enterprises face when trying to turn data into a reliable, governed currency.
How does the transition from data being a strategic asset to a structural foundation fundamentally change how enterprises approach their AI deployments?
For years, we treated data as a strategic reservoir—something you gathered in a lake or warehouse and hoped to derive insights from eventually. But as Salesforce finalized its acquisition of Informatica in November 2025, the industry realized that data is no longer just “strategic”; it is the literal scaffolding of the entire enterprise. When you move into production with AI, you aren’t just looking for a cool chatbot; you are building an infrastructure where the AI is expected to pull from every corner of the business. If that foundation is cracked or fragmented, the AI doesn’t just fail; it creates a massive operational risk that ripples through the organization. We are seeing a shift where the “heavy lifting” is moving behind the scenes into governance and quality, ensuring that when an agent reaches for a piece of information, it is touching something solid and reliable rather than a ghost of a deleted record. It feels like moving from a world where you are drawing blueprints to one where you are finally pouring the concrete, and you realize the quality of that concrete determines if the building stands or falls.
What does the concept of “trusted context” mean in a practical sense for an AI agent, and why is it described as the new currency for business?
Trusted context is the difference between an AI that sounds smart and an AI that acts like a tenured employee who has been with the company for twenty years. It is the connected, governed view of customers, products, and suppliers that allows an agent to understand the nuance of a business relationship. Without it, an AI is just a commoditized model; with it, the agent has the “currency” to make high-value decisions that actually move the needle. Think of the sensory relief of a customer service agent having a unified view of a client’s history across 63,000 different restaurant locations, as we see in large-scale operations. When an agent has this context, it doesn’t just provide an answer; it provides a business-relevant action that respects lineage, security, and ownership. We are entering an era where the model itself is a bolt-on interface, while the true differentiation lies in how well that agent understands the unique, gristly reality of the enterprise it serves.
Why do you believe that agentic AI systems are uniquely positioned to expose the long-standing failures of master data management?
Agentic systems are like a bright spotlight shined into a dark, cluttered basement; the moment they start interacting with enterprise processes, they instantly find the duplicate records and inconsistent definitions we’ve ignored for a decade. Recent research showed that a staggering 84% of data leaders believe their organizations need significant strategy overhauls before AI can truly succeed at scale. This isn’t just a technical glitch; it’s a visceral realization that our data governance has often been an afterthought. When an AI agent tries to execute a cross-sell but sees three different versions of the same customer, the system stutters, and the failure is immediate and visible. Master data management (MDM) used to be the overlooked infrastructure of the IT department, but now it’s the frontline of the AI revolution because you cannot have an autonomous agent navigating a landscape of broken data links.
With Salesforce integrating Informatica into its broader ecosystem, how do you see the competitive landscape with players like Microsoft, Snowflake, and Databricks shifting?
The battlefield has moved away from simple data storage and basic analytics toward who can provide the most robust “trusted layer.” Microsoft is aggressively building out its Fabric and OneLake ecosystems, coupled with Purview for governance, while Snowflake is rapidly expanding its semantic and catalog capabilities to keep pace. Even in the niche markets, we see massive moves, such as SAP acquiring Reltio in May 2026 to bolster its graph-oriented architecture and API-first design. It’s a high-stakes race where everyone—from Oracle to Databricks—is trying to create a consistent layer of business context that can be used across every application and workflow. You can feel the tension in the market as these giants realize that being the place where data “lives” is no longer enough; they must be the place where data is “governed” and “understood.”
What can we learn from the implementation experiences of major organizations like Yum Brands and TELUS regarding the real-world value of unified data?
The success stories of Yum Brands and TELUS provide a grounded, sensory look at what happens when you finally get the “plumbing” right. Yum Brands, which manages a massive global footprint of over 63,000 restaurants including KFC and Taco Bell, found that their teams were wasting an incredible amount of energy just cleansing location data before it could even be used. By implementing Informatica MDM, they turned that chaotic friction into a streamlined process where location data became a central, reliable asset for the whole business. Similarly, TELUS used these tools to bridge the gap between their telecommunications and health services, creating a unified customer view that allowed for much more targeted and effective marketing. These aren’t just wins on a spreadsheet; they represent a fundamental shift in how employees interact with information, replacing the frustration of “which version is correct?” with the confidence of a single, trusted source.
In the context of scaling AI, why is “headless data management” becoming a critical requirement for organizations?
Headless data management is about stripping away the need for custom, clunky integrations every time you want to launch a new use case or agent. Organizations are tired of the “integration tax” that slows down innovation, so they are looking for ways to let agents and applications access governed services directly through protocols like the Model Context Protocol. It allows an AI system to dip into the enterprise data pool, grab what it needs with full governance and lineage intact, and move on without a developer having to build a bespoke bridge. This approach creates a more fluid, organic environment where data services are “exposed” rather than “locked,” making the entire AI ecosystem much more agile. When executed well, it feels less like a series of rigid pipes and more like a pervasive atmosphere of intelligence that any system can breathe in.
What are the primary reasons many data programs still struggle despite high executive interest, and how can leaders avoid these pitfalls?
The most common tragedy I see is that technology decisions almost always move faster than the governance models required to support them. Organizations rush to buy the latest AI or MDM tool, but they fail to establish clear ownership, stewardship, or accountability, which means the tool is essentially sitting on a foundation of sand. Many data programs struggle because governance is seen as a “later” problem, or because different business units are allowed to maintain competing definitions of the same core data. To avoid this, leaders must reverse their approach: start with the business priorities, translate those into a data strategy, and only then pick the technology. It’s about having the organizational discipline to define who “owns” the customer record before you spend millions of dollars on an AI that is supposed to talk to that customer.
What is your forecast for the evolution of the enterprise data layer over the next few years?
I believe we are heading toward a period where the “winner” in the AI race won’t be the company with the most sophisticated large language model, but the one with the cleanest, best-governed data foundation. We will see the “Agent Fabric” become a standard part of the corporate architecture, where governance and quality controls are embedded directly into the AI’s decision-making loops rather than being a separate check-point. By 2026, I expect to see audited outcomes where companies can prove that their AI deployments are generating measurable value because they’ve finally solved the master data problem. The “trusted context” we’ve discussed will become so foundational that we’ll stop talking about it as a separate initiative and simply view it as the way modern business operates. My advice for readers is to stop looking for the “magic model” and start investing in the gritty, unglamorous work of data stewardship today, because that is where your future competitive advantage is currently hiding.