Why utilities need an information foundation for the AI era 

Utilities are navigating one of the most significant periods of transformation in their history. Grid modernization, renewable integration, electrification, workforce turnover, and increasing energy demand are creating new operational and financial pressures. At the same time, organizations are exploring how AI assistants, copilots, and autonomous agents can help improve productivity, accelerate decision-making, and streamline business processes. 

As SAP customers look to operationalize AI, one reality is becoming clear: success is determined less by the AI model itself and more by the information foundation beneath it. Every AI agent ultimately needs trusted information to make decisions, governed authority to take action, and secure mechanisms to exchange information across organizational boundaries. 

This is where OpenText™ Solutions for SAP play a critical role. OpenText helps organizations connect content, processes, records, and engineering information directly to SAP business transactions, creating a governed information foundation that enables AI to scale safely and effectively across the enterprise. 

As utilities move from AI experimentation to production deployment, three capabilities become essential: Decide, Act, and Transact. These capabilities define what every AI agent requires to operate within a regulated, mission-critical environment. 

Decide: Read trusted context 

Before an AI agent can make a recommendation, it must first understand the situation it is evaluating. That requires access to trusted context. 

In utility operations, decisions depend on engineering drawings, maintenance procedures, inspection reports, work histories, asset documentation, compliance records, operational reports, customer bills and countless other sources of information. Yet these records often reside across multiple systems and repositories, making it difficult to ensure employees—and AI agents—are working from complete and current information. 

The challenge becomes even more important as organizations deploy AI. Agents can only make decisions based on the information available to them. If information is incomplete, outdated, or disconnected from business processes, the result can be inaccurate recommendations delivered with unwarranted confidence. 

The answer is governed content. 

The agent reads governed asset and operational records that are current, complete, and properly organized according to business, regulatory, and operational requirements. 

OpenText™ Content Management and OpenText™ Documentum™ Content Management (CM) provide that foundation by connecting enterprise content directly to SAP business processes. Utilities can create a trusted information environment where documents, records, engineering information, and operational content remain governed throughout their lifecycle. 

When AI agents access content through a governed information layer, they are working from approved versions, validated records, and controlled information sources. Maintenance teams gain access to current procedures. Engineers work from approved documentation. Field service personnel can quickly locate the information required to support critical decisions. 

In an AI-enabled utility, trusted information is not optional. It is the foundation that enables every subsequent decision. 

Act: Ensure AI operates within authorized boundaries 

Once an agent can access trusted information, the next question becomes whether it should act on that information. 

For utilities, governance is the difference between an interesting AI pilot and a production-ready AI capability. Every organization must answer a simple but critical question: What is the agent allowed to do? 

The principle is straightforward:  

The agent does only what you granted—and every action is logged. 

Utility business processes require approval controls, segregation of duties, auditability, compliance oversight, and documented accountability. AI cannot operate outside these requirements. Instead, it must function within the same governance framework that guides human workers. 

A strong example is OpenText™ Vendor Invoice Management for SAP® Solutions.  Invoice processing is a business process where organizations want increased speed and efficiency while maintaining control. Intelligent automation can classify invoices, extract information, validate data, route exceptions, and support workflow execution. However, the system remains governed by established approval processes, business rules, authorization controls, and audit requirements. 

Every action remains visible. Every exception remains traceable. Every approval is documented. 

Transact: Give AI a transaction passport 

Many of the highest-value utility processes extend beyond company boundaries. Capital projects involve engineering firms and contractors. Procurement processes involve suppliers and manufacturers. Commercial operations involve customers and external partners. 

As AI evolves, agents will increasingly participate in these interactions. 

To do that, they need what OpenText describes as a transaction passport—the ability to exchange information across organizational boundaries using identity, trust, and verifiable evidence.  

A transaction passport answers several critical questions: 

• Who is participating in the transaction? 

• What information is being exchanged? 

• Can the information be trusted? 

• Is there a verifiable record of what occurred? 

• Can the organization prove compliance later if challenged? 

For utilities, one of the clearest examples is engineering collaboration during capital projects. 

Engineering drawings, specifications, design documents, redlines, and technical packages routinely move between utilities, EPC firms, contractors, and suppliers. Without governance, these exchanges can create delays, version-control problems, rework, and compliance risks. 

Solutions such as OpenText Content™ Management for Engineering help address this challenge through capabilities including transmittal and submittal management. These processes support the controlled exchange of project documentation while maintaining traceability, version control, review workflows, approvals, and audit trails. 

When an engineering document is shared, stakeholders can verify which version was sent, who reviewed it, what changes were requested, when approvals occurred, and what documentation supports the final decision. The entire exchange becomes governed and defensible. 

This same principle applies to the future of agentic AI. Before an AI agent can safely engage with suppliers, contractors, partners, or commercial customers, it must have access to trusted identity, governed information, and verifiable evidence. In other words, it needs a transaction passport. 

What utilities need to support agentic AI 

Every effective AI agent depends on three capabilities. It must be able to read trusted context to decide. It must be able to operate within authorized boundaries to act. And it must possess a transaction passport to transact across organizational boundaries using identity, trust, and verifiable evidence. 

Through its portfolio of SAP-integrated information management solutions and its long-standing partnership with SAP^®, OpenText helps utilities build the governed foundation that connects content, records, engineering information, workflows, and business processes directly to SAP environments. 

Ready to learn how OpenText can help your utility establish an AI-ready information foundation for SAP? Speak to an expert.