Artificial intelligence (AI) is fundamentally changing how utilities interact with, and benefit from, the massive amounts of data that is at their disposal. Today, AI-powered solutions help provide visibility into the grid edge, deliver bill savings for customers, support conservation efforts, and much more. For AI to truly maximize outcomes for both customers and the business, a number of key tenets should be central to a utility’s AI strategy.

Better data for better outcomes

First, any progress, be it AI-powered or not, will be dependent on data that is secure, clean, connected, and cataloged. All too often, utilities struggle with data that is unreliable, cryptic, or too unwieldy to understand and leverage. Making matters more challenging, personally identifiable information (PII) and security concerns can obstruct access for employees who need the data. Further, data is still often kept in transactional systems or other databases that are disconnected and siloed away from each other. These situations slow or fully restrict innovation.

Freer-flowing data is imperative to producing better visibility, analysis, and smarter responses to an increasingly complex energy ecosystem. When data is responsibly and securely brought together, standardized, and prepared for AI, rapid progress can be made toward solving the myriad of challenges our industry faces.

To address this need, Oracle is developing the Energy and Water Data Exchange* to translate data and add context so that it can be readily understood by both humans and AI models. The solution provides contextual data and relationships for use by large language models (LLMs). Coupled with Oracle Energy and Water Data Intelligence, utilities can unify and streamline all their data in a single repository so that employees with predefined authorization can perform cross-domain business analytics and self-service data science discovery.

Build-in, not bolt-on

To achieve the best outcomes, utilities must commit to embedding AI in their core business processes and application instead of relying on a siloed data science capability. By integrating AI, utilities help ensure the freshest insights are in the hands of employees who are best positioned to act on them. Otherwise, organizations run the risk of disjointed analyses that create chaos through multiple, conflicting models that quickly can become abandoned investments.

In this respect, we practice what we preach. Oracle has spent the past 20 years building its AI foundations. We embed AI into our entire technology stack—from the database to our mission-critical applications. This holistic approach allows our customers to gain an advantage by ensuring that data-driven insights are omnipresent.

Today, this approach extends to GenAI, which is receiving prominent industry attention for its promise to help fuel next-level problem-solving. Oracle provides Gen AI-based applications and technology that allow organizations to better automate key business functions, improve decision-making, and enhance their customers’ experiences.

Put AI to best use

With the right data and AI strategies in place, utilities have the flexibility to identify and choose the business problems that can best be solved with AI. In our work with utility leaders, we’ve found that optimizing customer engagement is at the top of everyone’s list.

By leveraging AI capabilities in the Opower Customer Engagement platform, utilities can support limited-income customers, shift customer load, and increase overall digital engagement. For example, AI embedded in the platform removes the complexity of identifying customers struggling with energy burden, increasing effectiveness of targeted campaigns and programs. As a result, a North American IOU more than quadrupled enrollments in LIHEAP programs. Another utility used AI-powered disaggregation insights to shave 14.5 MW of peak load when it needed it most.

Additionally, utilities are using Oracle products to tap prebuilt, deep machine learning (ML) and advanced analytics that help solve complex use cases. This includes finding the literal needle in the haystack with trained algorithms. For example, Oracle customers use our prebuilt solutions to (1) pinpoint where EVs are charging, (2) identify water leaks before they impact the customer bill, or (3) predict storm damage. One North American IOU recovered over $10 million in revenue using Oracle theft and broken meter detection analytic models. Another North American utility identified potentially overloaded distribution assets (e.g., line transformers and feeders) and then marketed new Time of Use (TOU) rates to residential EV owners.

Looking ahead, several active AI use case pilots are focusing on enhancing the customer and call center agent experience. In these scenarios, AI, sentiment analysis, chatbot, and guided resolution tools automate and simplify tasks and reduce the workload, allowing call center agents to focus on the more complex customer needs.

Taking a data-driven, complete approach

Any conversation about AI must first consider the quality, completeness, and consistency of the data being fed into any model. For utilities to achieve the most meaningful outcomes, data must be aggregated, qualified, and intelligible to AI platforms; this is not a trivial endeavor.

AI’s potential is truly great and untapped, but it’s not a silver bullet. AI is part of a broader data and business transformation story. When AI is built into end-to-end business processes, utilities will accelerate and maximize benefits for their employees, their customers, and the broader communities they serve.