Anand Naidu stands at the intersection of complex backend architecture and intuitive frontend delivery, bringing a wealth of experience in building the connective tissue that powers modern enterprise applications. As a veteran developer proficient in navigating the nuances of disparate coding languages, he has witnessed firsthand how the most sophisticated AI models can stumble when fed a diet of fragmented or low-quality data. In this discussion, we explore the strategic implications of SAP’s recent move to acquire Reltio, a shift that signals a fundamental change in how corporations view data management. Our conversation delves into the necessity of moving beyond isolated data silos toward a “system of context,” the technical challenges of creating a “golden record” in a multi-vendor ecosystem, and the critical role of real-time entity resolution in empowering autonomous, agentic AI to make high-stakes business decisions with confidence.
Why is the industry shifting from a model-centric AI approach to one focused on data unification? How does establishing a “system of context” ensure that autonomous agents make reliable decisions instead of operating in isolated silos? Please share specific examples or metrics that illustrate this transition.
For a long time, the tech world was obsessed with the “brain” of the operation—the AI model itself—but we’ve realized that even the smartest brain is useless if it’s hallucinating based on bad information. We are seeing a massive pivot because fragmented data across various business units remains the primary barrier to delivering reliable outcomes, often stalling AI initiatives before they can even scale. By establishing a “system of context,” we are essentially creating a unified layer that connects structured and unstructured data across different applications, ensuring that an autonomous agent isn’t just looking at a single spreadsheet but understands the entire history of a customer or product. This transition is clearly reflected in SAP’s strategy to integrate Reltio into its Business Data Cloud, moving away from standalone AI capabilities toward a foundation where data quality is the main priority. Without this context, agents operate in silos, leading to “model-centric” failures where the AI lacks the situational awareness needed to trigger a meaningful business action.
Integrating internal data with third-party sources remains a major hurdle for most enterprises. What specific steps are required to build a “golden record” across disparate systems, and how does this approach reduce the high costs typically associated with maintaining complex data pipelines?
Building a “golden record” is a meticulous process that involves advanced entity resolution to identify and merge duplicate records of customers, products, and suppliers across both SAP and non-SAP systems. It starts with data cleansing and harmonization, where we strip away the noise and inconsistencies that naturally occur when different departments use different software. By centralizing this through a cloud-native master data management platform, organizations can finally move away from the “spaghetti code” of custom-built integrations that are notoriously brittle and expensive to maintain. This architectural shift significantly reduces the high costs of maintaining complex data pipelines because you are governing the data at the source rather than trying to fix it every time it moves between systems. It transforms data management from a high-maintenance back-office chore into a strategic, streamlined asset that serves the entire enterprise ecosystem.
Organizations often struggle with inconsistent data regarding suppliers, customers, and products. How does real-time entity resolution transform procurement workflows, and what specific metrics should leaders track to measure the impact of improved data trust on reducing operational friction?
In a typical procurement workflow, a lack of data trust can lead to massive delays or even financial loss if an agent doesn’t realize two different vendor entries actually represent the same entity with a high-risk profile. Real-time entity resolution changes this by providing a singular, trusted view that allows AI agents to assess supplier risk on the fly and trigger immediate actions based on accurate information. Leaders should keep a close eye on metrics like the time-to-resolution for data discrepancies and the reduction in manual overrides required by procurement officers. When you reduce operational friction through trusted data, you see a direct improvement in the reliability of both analytical outputs and day-to-day operational decisions. It’s about moving from a state of constant manual verification to a high-velocity environment where the data foundation is solid enough to support automation without human hand-holding.
As agentic AI becomes more prevalent, these systems must trigger actions based on real-time, trusted information. What are the primary risks of deploying autonomous assistants without a unified data foundation, and how can technical teams effectively bridge the gap between fragmented data environments?
The primary risk of deploying autonomous assistants like Joule in a fragmented environment is “informed inaccuracy,” where the AI executes a transaction or a legal commitment based on an incomplete or outdated dataset. If your agentic AI triggers a procurement order based on a supplier record that hasn’t been updated with recent performance failures, the cost of that error can be astronomical. Technical teams can bridge this gap by implementing a centralized management layer that acts as a “single source of truth” across heterogeneous environments, ensuring that every agent is pulling from the same pool of harmonized data. This requires a shift toward an “AI-first” and “suite-first” strategy, where interoperability between SAP-native data and third-party sources is treated as a core requirement rather than an afterthought. By embedding data governance directly into the AI strategy, teams can mitigate the risk of autonomous “hallucinations” and ensure that every action taken by the system is defensible and accurate.
Enterprise ecosystems are rarely limited to a single software vendor. How can companies maintain data consistency when operating across multiple cloud platforms, and what role does a centralized management layer play in supporting complex, multi-agent workflows that require low latency?
Modern enterprises are almost always multi-vendor landscapes, which makes the challenge of data consistency a cross-platform struggle that can’t be solved within a single walled garden. To maintain consistency, companies need a centralized management layer that provides a unified data foundation, allowing information to flow seamlessly between SAP and non-SAP environments without losing its context or integrity. This layer is crucial for multi-agent workflows because it ensures that when one agent passes a task to another across different cloud platforms, the underlying data remains synchronized and available with minimal latency. We are looking at a future where real-time data availability is the benchmark for success, and having a cloud-native MDM capability integrated into the core business data cloud is the only way to support these high-speed, complex interactions. It ultimately allows organizations to scale their AI initiatives across diverse software stacks without the typical “integration tax” that slows down digital transformation.
What is your forecast for the future of enterprise AI and master data management?
I forecast that master data management will shed its reputation as a back-office discipline and become the most critical strategic pillar for any organization looking to scale AI. By the time SAP’s acquisition of Reltio is finalized in the second half of 2026, we will see a market where standalone AI tools are viewed as secondary to the integrated data platforms that feed them. We are moving toward a reality where “AI-readiness” is defined entirely by the quality of a company’s data foundation and its ability to achieve real-time, cross-platform interoperability. This means that vendors will continue to swallow up specialized data governance players to ensure their AI agents have the trusted context they need to function autonomously. Ultimately, the winners in this space will be the companies that treat data as a living, breathing asset that is harmonized, governed, and ready for action in every corner of the business.
