Utilities generate enormous volumes of valuable data. Yet answering a seemingly straightforward operational question can still require several systems, multiple teams and hours of manual investigation.
Why is a settlement run incomplete? Does the consumption profile in SAP match the data held by the metering system? Which customers exhibit the greatest flexibility during peak-price periods?
The information needed to answer these questions often exists but it is distributed across SAP, metering infrastructure, energy data platforms and market communication systems. Finding and interpreting it requires both access to the data and a deep understanding of utility processes.
exnaton is working to make this expertise available directly where many utility teams already work: SAP Joule.
exnaton has developed an integration that enables SAP Joule to communicate with exnaton’s specialized energy agents. Utility users can ask questions in natural language through Joule and receive insights from exnaton’s domain-specific agents in the same interface, alongside SAP's own agents.
SAP Joule is SAP’s AI copilot and increasingly serves as a natural-language interface through which users interact with SAP systems and business processes.
The exnaton integration makes specialized energy and utility intelligence accessible through that same interface. A connector agent allows Joule to call exnaton’s energy agents using Agent2Agent, or A2A, communication on SAP BTP.
A2A is an open protocol for communication between AI agents. SAP uses it to enable Joule to interact with externally developed, code-based agents, leveraging the power of SAP BTP as underlying system.
In practice, that means:
Utility teams therefore gain access to exnaton’s domain intelligence alongside SAP’s own agents without an additional tool, separate login or replacement of their core SAP system.
The first integration demonstrates two exnaton specialists, each addressing a common operational challenge for utilities.
Settlement problems are rarely contained within one system. When a process fails or remains incomplete, employees may need to compare metering information, customer master data and market communication records manually.
The exnaton Data Quality Agent supports this investigation across connected systems. It can:
Instead of navigating several applications and coordinating investigations across teams, users can begin with a natural-language question in Joule.
This can help utilities shorten resolution times, reduce manual investigation and give operations teams a clearer understanding of data-quality problems.
Utilities possess increasingly granular energy data, but turning that data into a useful answer often requires specialist analysis.
The exnaton Energy Intelligence Agent enables teams to ask questions about consumption, feed-in and metering data using natural language. It can help users investigate:
This makes granular energy information more accessible to business and operational users who need answers but may not work directly with databases or analytics tools.
Utility processes involve highly specialized data models, regulations and operational dependencies. An incomplete settlement run, for example, may originate in metering data, master data, market communication or the interaction between several systems.
A general-purpose AI assistant may understand the question but lack the industry context and regulatory knowledge needed to determine:
This is where specialized energy agents create value. They combine AI-based interaction with knowledge of utility data, processes and energy-market requirements.
For utilities whose system of record is SAP, connecting this domain intelligence to Joule brings the two essential elements together: enterprise context and utility-specific expertise.
Every action that leaves the system passes a human approval gate. That single design choice does two things at once: it keeps a person accountable for every consequential step, and it keeps these agents in the EU AI Act's limited-risk tier, with the transparency obligations that tier requires rather than the heavier conformity regime that applies to high-risk systems. For a utility evaluating AI in a regulated industry, that's not a footnote, it's the reason this can move from pilot to production without opening a new compliance track.
By bringing exnaton’s specialized energy agents into SAP Joule, utilities can obtain consolidated answers across fragmented system landscapes without manually comparing records, tracing inconsistencies or analysing granular time-series data. Product, customer service and operations teams can access utility-specific insights through natural-language questions in a familiar SAP interface, while specialists remain in control of operational decisions.
The integration adds AI-powered capabilities without replacing SAP or rebuilding the existing system architecture. Its modular design also provides a scalable foundation: utilities can expand the available functionality by adding specialized agents for new use cases rather than developing a separate integration each time.
Talk to our team about how exnaton’s energy agents can support your utility’s data analysis, settlement and operational workflows through SAP Joule. Talk to our experts.
The integration enables SAP Joule to communicate with exnaton’s specialized utility agents through the open Agent2Agent protocol. Users can ask utility-specific questions in Joule and receive answers based on data and domain logic handled by exnaton’s agents.
Agent2Agent, or A2A, is an open protocol that allows AI agents built with different technologies to communicate, exchange information and delegate tasks. SAP supports A2A for connecting externally developed agents with Joule.
The integration is designed to make exnaton’s agents accessible through SAP Joule on SAP BTP. Users can therefore interact with the agents from a familiar interface rather than moving to a separate AI application.
No. exnaton adds specialized utility intelligence to the existing SAP environment. SAP remains the core system of record.
The demonstrated architecture requires human approval before an action is sent to or executed in an external system. This keeps utility specialists in control of consequential operational actions.
Yes. The modular A2A architecture allows new specialists to be added without creating a completely separate Joule integration for every use case.
