Managing data quality, utility operations, and the growing role of AI were central themes in my recent conversation with SAPinsider community member Atul Joshi. During our discussion, Atul offered practical, real‑world examples from utility companies navigating the transition from traditional processes to modern, data‑driven operations. Atul’s perspective underscores a broader trend we continue to see among SAP customers. Organizations want AI, automation, and modernization but are still working to clarify what they want these technologies to achieve and how they will support long‑term operational outcomes.

The Foundation: Data Quality Still Determines Everything

Atul emphasized that a ‘clean core’ strategy - starting with clean, complete master data is the foundation for every modernization initiative. As energy organizations modernize data models, collection, and management must reflect real‑world operational complexity. For example, he highlighted how customer data requirements must include fields that separate commercial from residential customers – allowing a representative to differentiate between a residential customer that requires a phone number and email address and commercial entities where additional information fields are important ((such as business type, tax identification, load profile category  Field requirements may also differ by communication patterns. Atul shares that he’s seen clients missing even basic data. Data management in the energy industry can also face challenging operational realities from the field. Meter readers depend on precise location data, such as latitude and longitude, to access meters safely and efficiently. He’s seen situations where energy organizations face obstacles like fencing or even dogs, which can make it challenging to get the data needed. He notes that before organizations start testing AI use cases, an essential element is data which includes timely and reliable collection, as well as structure. But Atul’s point extends beyond basic completeness to architectural flexibility. By leveraging SAP BTP as an extension layer, utilities can capture field-specific data—like GPS coordinates for meters or 'restricted access' notes—without cluttering the S/4HANA core. This side-by-side extensibility ensures the core remains standard while the business logic adapts to field reality

AI in Utilities: Big Aspirations and Building Expectations

Before implementing AI, utilities must understand not only the desired outcomes but also the data inputs, governance requirements, and operational implications. When it comes to AI, Atul noted there is strong enthusiasm across utility companies. A lot of conversation in the media is around the electrical grids needed to support AI. But there is more to the story, many utilities are already exploring tools like SAP Joule and the AI foundation on SAP BTP. Atul specifically noted that the next evolution involves deploying Joule as an autonomous agent. This isn't just about a chatbot; it's about an agent capable of navigating complex utility business objects on BTP to automate service reconnections or billing adjustments without manual intervention. Yet, as he observed, leaders often begin with the desire to “do AI” without defining what outcomes they expect. Atul put it succinctly: “Everyone wants AI, but what they want from AI—nobody knows.” This gap between aspiration and clarity is one of the most important challenges he sees across the industry. Companies view AI as a transformative force but often conflate it with chatbots or simple query‑response tools. As he noted, AI is not merely a conversational interface, it is an analytical engine that depends entirely on the quality, completeness, and structure of the data it is fed. Atul goes on to explain that AI’s value is determined by the data you provide bad data leads to bad insights, regardless of the sophistication of the model. This reality is particularly significant in the utility sector, where customer information, usage data, location intelligence, and compliance records all carry regulatory and operational implications. Utility companies cannot afford unreliable predictions or incorrect recommendations. Another theme Atul raised is data sensitivity. Utilities hold extensive customer information, and many organizations are understandably cautious about what they share with AI systems. He described this as being “conscious” rather than fearful which reflects a growing maturity across industries and countries about the importance of data governance in AI‑enabled environments. Listen to a clip from our interview here:

What This Means for SAPinsiders

For SAPinsider members, Atul’s insights reinforce several core imperatives: Data Governance Must Match Operational Reality: From customer master data to field service notes, utilities must align data policies with real‑world workflows. Misalignment creates gaps that become barriers to automation and AI. AI Readiness Starts with Clear Use Cases: Before deploying tools like SAP Joule or training models, organizations need documented, prioritized scenarios where AI can directly impact safety, accuracy, cost, or customer experience. Sensitive Data Requires Thoughtful Guardrails: Utilities’ concerns about data sharing reflect the importance of clearly defining governance around AI within the organization. Community Knowledge is a Strategic Asset: Atul’s frontline examples show the value of shared lived experience in developing practical approaches to digital transformation and AI adoption.