# Utilities: keep essential services working
The estate of an electricity distributor and a water utility on the fourteen layers. Four worked examples placed on it, who uses AI today, and the rules that bind the work.
Author: Murali Sid (https://linkedin.com/in/muralisid)
Source: https://www.agenticarchitectureskills.com/use-cases/utilities
Published: 2026-09-18
Updated: 2026-10-15
Licence: CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0/
Cite as: Murali Sid. Utilities: keep essential services working. Agentic Architecture Skills. https://www.agenticarchitectureskills.com/use-cases/utilities
---
Utilities keep power and water reaching homes and businesses. They run networks, read meters, bill customers, fix faults and answer to regulators. This page maps that work onto the guide's frame. The frame is the [layers you own](/layers), a [lifecycle you run](/architecture/agent-lifecycle), and the [capabilities you add](/architecture/system-view). It shows the fourteen layers of a typical utility estate and the systems in each. It places four worked examples on those layers. The short answer: agents fit around the systems utilities already run. In the deployments found, language-model systems advise and people decide.
## Who owns which work [#who-owns-which-work]
Ownership decides which use case belongs to whom. It is not the same in electricity and water.
* **Electricity retailers bill households.** In Australia, an electricity distributor does not bill households. Retailers do. Billing exceptions, payment plans and hardship belong to the retailer.
* **Distributors run the network.** A distributor owns poles, wires, substations and transformers. It runs the control room, connects customers and restores supply. It bills retailers for use of its network.
* **Meters are split too.** In most Australian states, household smart meters belong to competing metering coordinators. Victoria is the exception. Its distributors own the household smart meters, so they run the meter head-end and meter data.
* **New Zealand has the same split.** Retailers, called traders, bill customers. Distributors bill retailers for network use. Market parties exchange standard files called EIEPs through the Electricity Authority's registry.
* **Water utilities bill households directly.** A water utility treats and delivers water and collects wastewater. It bills households itself, usually every quarter. There is no wholesale market in between.
| Work | Electricity in Australia and New Zealand | Water |
| --------------------------------------- | ----------------------------------------------------------------------------------- | --------------------------------------------------------- |
| Customer bills, payment plans, hardship | The retailer | The water utility |
| Network, outages, connections | The distributor | The water utility |
| Household meters | A metering coordinator. The distributor in Victoria | The water utility. Most meters are still read by a person |
| Market messages | AEMO's B2B e-Hub and MSATS in Australia. The registry and EIEP files in New Zealand | None |
AEMO is the Australian Energy Market Operator. Its business-to-business (B2B) e-Hub carries service orders, meter data and notices between market parties. Its market settlement and transfer system (MSATS) holds the standing data for each meter. The Australian Energy Regulator (AER) enforces the retail rules.
## The estate [#the-estate]
These are the systems a distributor of about one million customers and a metropolitan water utility typically run. Vendors in brackets are examples, not a claim about any named utility. OT means operational technology: the control systems that run the physical network. SCADA is the main one. It is the control-room system that monitors and operates field equipment.
| Layer | Electricity distributor | Water utility |
| ---------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| [Infrastructure and compute (R01)](/layers/r01-infrastructure) | Two data centres plus public cloud. A separate OT network for the control centre and substations. Mesh radio or cellular for smart meters | The same pattern. Telemetry to pumps, reservoirs and plants. Low-power networks for sensors and digital meters |
| [Data platform (R02)](/layers/r02-data-platform) | Warehouse or lakehouse (Snowflake, Databricks, Fabric, SAP Datasphere). BI (Power BI). A historian copy. Data marts for reports to the AER | The same, plus lab data and hydraulic model outputs |
| [Integration fabric (R03)](/layers/r03-integration-fabric) | Integration bus or platform (SAP Integration Suite, MuleSoft, Boomi). API gateway. File transfer. A market gateway to the B2B e-Hub. A one-way OT to IT path through a buffer zone (DMZ) | The same, without the market gateway. An IoT platform for sensors |
| [Systems of record (R04)](/layers/r04-systems-of-record) | Network billing to retailers. Meter data management where it owns meters (Itron, Oracle, Siemens). Asset system (SAP PM, IBM Maximo, Oracle WAM). GIS, the network map (Esri, GE Smallworld). ERP. Records. Outside records: MSATS or the NZ registry | Customer billing system, the CIS (SAP, Hansen, Oracle, TechnologyOne). Asset system. GIS. ERP. Records. Lab system (LIMS) |
| [Line of business and OT (R05)](/layers/r05-lob-and-ot) | SCADA, distribution management (ADMS) and outage management (OMS) (GE Vernova, Schneider, Siemens, Oracle). Smart meter head-end (Itron, Landis+Gyr). Historian (AVEVA PI). Field mobility. Vegetation platforms. Contact centre telephony | SCADA (Schneider, AVEVA, Citect). Treatment plant control. Historian. Event analytics (TaKaDu, StormHarvester). Hydraulic models. Leak sensors. Digital meter head-end. Field mobility |
| [Intelligence and learning (R06)](/layers/r06-intelligence-and-learning) | Data science platforms (Databricks, Azure ML). Models inside packages, such as SAP's release confidence on held reads. Load forecasting. Image analytics as a service (Tapestry, Neara, AiDash, LiveEO). Microsoft 365 Copilot | The same, plus anomaly detection on telemetry and acoustic leak models |
| [Agent platform (R07)](/layers/r07-agent-platform) | Mostly new. Today: robotic process automation (UiPath, Power Automate) and rule engines (SAP BPEM, ServiceNow, Pega). Emerging: SAP Joule Studio, Copilot Studio, Agentforce, the Maximo MCP server | The same, plus customer chatbots |
| [Productivity and collaboration (R08)](/layers/r08-productivity-and-collaboration) | Microsoft 365. Spreadsheets hold many exception trackers. Tablets in the field | The same |
| [Experience and channels (R09)](/layers/r09-experience-and-channels) | Outage map and SMS alerts. Connections portal. Fault and life support lines. Retailers run the bill portal, app and chat | Customer portal and app. Contact centre. SMS and web notices for outages |
| [Security and identity (R10)](/layers/r10-security-and-identity) | Workforce identity (Entra ID, Okta). Privileged access (CyberArk). Security monitoring (Splunk, Sentinel). OT monitoring (Claroty, Dragos, Nozomi) | The same, plus customer identity |
| [Governance, risk and sovereignty (R11)](/layers/r11-governance-risk-sovereignty) | Risk and compliance tool (ServiceNow IRM, Archer, Protecht). Obligations register. Reports to the AER, or the Commerce Commission and Electricity Authority in New Zealand | Risk and compliance tool and obligations register. Reports to the economic, health and environment regulators. Taumata Arowai in New Zealand |
| [Observability and FinOps (R12)](/layers/r12-observability-and-finops) | Application and interface monitoring (Dynatrace, Splunk, SAP Cloud ALM). Market message monitoring. Cloud cost tools. Almost no tracing of model calls | The same |
| [Operating model and supervision (R13)](/layers/r13-operating-model) | IT service management (ServiceNow). Control room shifts with authorisation levels. Contractor management. No agent supervisor role yet | The same, plus treatment plant operators and quality assurance (QA) in billing and the contact centre |
| [Agent data engineering (R14)](/layers/r14-agent-data-engineering) | Mostly new. Data catalogs (Purview, Collibra). Master data governance (SAP MDG). BI semantic models. A few retrieval pilots over manuals | The same, plus contact centre knowledge bases |
Five points on this estate matter most for agents.
* **The OT network stays apart.** Control-system data usually may not leave it. That decides where an agent runs and what it reads. The historian is the safe copy for an agent to read.
* **Three records must agree.** In electricity they are the GIS network model, the asset register and the market's standing data. They often do not, and that gap drives exceptions in every use case below.
* **The warehouse is a good source for history.** It is a poor source for the current state, because it loads overnight (typical, unsourced).
* **Interface monitoring shows that a message moved.** It does not show that the business outcome happened. Tracing of model and tool calls is mostly absent.
* **Supervision already exists.** Billing teams sample released work. Contact centres score calls. Control rooms have authorisation levels. The missing role is a person who owns an agent's results and its changes.
## The use cases on this estate [#the-use-cases-on-this-estate]
Four worked examples sit on this estate. Each one names its systems by layer, the process today, the process with agents, and who approves where.
| Worked example | Who owns it | Main layers | What the agent writes | Who approves |
| --------------------- | ------------------------------------- | ---------------------------------------------------------------- | --------------------------------------------------------- | ----------------------------------------------------------- |
| Billing exceptions | An energy retailer or a water utility | Systems of record, integration fabric, intelligence and learning | Proposals, case notes, draft market requests and messages | A fixed rule or a specialist. A trained person for hardship |
| Vegetation inspection | An electricity distributor | Line of business and OT, systems of record, integration fabric | A notification marked "agent proposed" | A planner, then a field assessor |
| Water network anomaly | A water utility | Line of business and OT, integration fabric, systems of record | An advisory view and draft notifications | The operator, the engineer, a named officer |
| AI use inventory | Any utility, for a regulator | Governance, risk and sovereignty, agent platform, observability | Draft inventory rows from the registry | A named person attests |
An energy retailer clears held reads and bills. A fixed rule or a specialist approves every change.
A distributor turns an image finding into a proposed notification that a planner approves.
An agent reads a copy of the alarm data and proposes a view. The operator decides.
List every AI use, the decision it shapes, and the safeguards around it.
### Customer journeys [#customer-journeys]
The customer account belongs to the retailer in electricity and to the utility in water. A customer calls, chats or uses the app. Common reasons are a home move, a meter read, a bill question, a payment plan, or trouble paying. Today a service representative works across the CIS, the CRM and the market gateway.
With agents, the agent says it is AI and offers a person. It checks identity, with a one-time code before any change. It drafts home moves through the existing market gateway. A fixed rule records life support at once. The agent explains bills without changing the account. It drafts payment plans within policy, and a trained person approves each one. It flags hardship and hands the customer over warm, with a summary.
Containment counts contacts that never reached a person. It does not count problems solved. AGL's assistant on IBM watsonx contains 36 percent of about 125,000 chats a month (IBM, undated) \[vendor]. PolyAI reports 67 percent containment for PG\&E's voice agent (2026) \[vendor]. The two count different things and are not comparable.
Kraken launched agents with Sierra in June 2026. They run set journeys such as identity checks, home moves and meter reads. One unnamed utility went live on about 1.3 million accounts \[vendor]. No results are published, and nothing was found on hardship or payment plans. Origin says its voice agent serves over 100,000 customers (12 February 2026, search excerpt). Hansen claims its AI agent can arrange payment plans \[vendor]. That is a product claim, not a named utility letting an agent change a plan alone.
## Named deployments [#named-deployments]
The map places each named deployment by area and by autonomy level. The levels come from the [autonomy contract](/architecture/autonomy-contract). They run from assisted (A1) through delegated tasks (A2) to supervised autonomy (A3). Most deployments advise at delegated tasks (A2). Kraken's set journeys act at one utility. Nothing was found at managed autonomy (A4).
Essential Energy in New South Wales was preparing an AI that reads crew notes for safety signals. It projects finding signals in 76 percent of cases, against 59 percent for its current rules (iTnews, 16 April 2026). That is a projection, not a measurement. Prefer a utility's own filing to a vendor's claim. Kraken claims A$170 million saved at Origin \[vendor]. Origin's half-year deck says A$100 to 150 million, seen only in a search excerpt.
No utility was found using agents for regulatory reporting or compliance evidence. Ofwat says regulatory reporting is one area where UK water companies report some AI use. It names no case (Kurrant, 18 June 2026).
## The rules that bind the industry [#the-rules-that-bind-the-industry]
**The hardship precedent.** In AER v Origin (Federal Court, 6 July 2022), automated processes changed payment plans without checking capacity to pay. Origin admitted over 100,000 breaches affecting over 90,000 customers. It paid an A$17 million penalty. This was automation, not AI. It still sets the line: a trained person decides any change to a payment plan. That is the rule for decisions about people. Red Energy paid A$474,600 in February 2025 over alleged failures to help customers in payment difficulty (search excerpt).
**The AEMC rule change of 10 September 2026.** The Australian Energy Market Commission (AEMC) made a final rule on that date. It strengthens payment difficulty and disconnection protections from 30 March 2027. It covers four of the five requests the AER lodged on 19 December 2025. A second rule, on payment plan standards, was still open on 15 October 2026. It would make retailers weigh what customers say they can afford before asking for documents. Feedback on a draft rule was due by 22 October 2026.
**The information path rule for SCADA.** Control systems act on the network. Agents stay on the information path around them. They read the historian copy, not the control system, and have no write path to OT. No utility was found letting an AI agent operate switching or protection (October 2026). This puts control actions in the [safety actuation zone](/architecture/deterministic-zones#safety-actuation). Joint government guidance on AI in operational technology says the same (NSA, CISA and others, December 2025).
**Telling customers about AI.** Ofgem and the UK Energy Ombudsman set out two expectations on 5 February 2026. Firms should say when AI is in use. They should make it easy to reach a person.
**Regulators asking for an AI list.** The New York Public Service Commission opened Case 26-M-0552 on 17 September 2026. It asks named utilities to file an inventory of their AI uses. The Arizona Corporation Commission opened a similar inquiry on 24 March 2026. The [AI use inventory](/use-cases/ai-use-inventory) shows how to build one.
## Evidence and limits [#evidence-and-limits]
The estate tables are industry knowledge, not a survey. No Australian or New Zealand utility publishes its full system estate. AEMO's pages returned errors, so market procedure details come from search excerpts.
Almost every result on this page is reported by the vendor or the operator. Few are independently audited. Containment figures count different things and should not be compared across utilities.
The named deployments that act alone run fixed rules or set journeys. No language-model agent was found changing a bill, a payment plan or a control system on its own. This guide recommends keeping it that way until a utility can measure the result and recover from a wrong action.
## Next step [#next-step]
Start with the worked example closest to your work, such as [billing exceptions](/use-cases/billing-exceptions). Then [choose a first task](/use-cases/choose-where-to-start).
Explore this topic
All sources checked on 15 October 2026 unless a date says otherwise. "Search excerpt" means the page itself could not be opened.
**Estate and market.** The estate tables follow industry practice, cross-checked against these references.
* [IEC TR 62357-1, reference architecture for power system information exchange](https://webstore.iec.ch/en/publication/6918), 25 October 2012.
* [CEN-CENELEC-ETSI Smart Grid Reference Architecture](https://www.cencenelec.eu/media/CEN-CENELEC/AreasOfWork/CEN-CENELEC_Topics/Smart%20Grids%20and%20Meters/Smart%20Grids/reference_architecture_smartgrids.pdf), November 2012.
* [ISA-95 standard page](https://www.isa.org/standards-and-publications/isa-standards/isa-95-standard).
* [AEMO B2B procedures](https://www.aemo.com.au/energy-systems/electricity/national-electricity-market-nem/market-operations/retail-and-metering/business-to-business-procedures). Search excerpt.
* [Electricity Authority on EIEPs](https://ea.govt.nz/industry/retail/eieps). Search excerpt.
* [AEMC, accelerating smart meter deployment](https://aemc.gov.au/rule-changes/accelerating-smart-meter-deployment), final rule 28 November 2024. Legacy meters replaced by 1 December 2030. Search excerpt.
* [APQC Process Classification Framework for Utilities](https://www.apqc.org/resource-library/resource-listing/apqc-process-classification-framework-pcf-utilities-pdf-version) and [EPA Effective Utility Management](https://www.epa.gov/sustainable-water-infrastructure/things-local-officials-should-know-about-effective-water). Reviewed 18 September 2026.
**Rules and regulators.**
* [AER release on Origin](https://www.aer.gov.au/news-release/origin-penalised-17-million-for-customer-hardship-breaches), 6 July 2022, as reported by [Region](https://region.com.au/origin-fined-17-million-admits-breaching-hardship-obligations-100000-times/572811).
* [AEMC final determination on payment difficulty protections](https://www.aemc.gov.au/sites/default/files/2026-09/final_determination_-_streamlining_payment_difficulty_protections_rrc0075.pdf), 10 September 2026. Search excerpt.
* [AEMC rule change on payment difficulty assistance standards](https://www.aemc.gov.au/rule-changes/strengthening-standards-payment-difficulty-assistance), consultation paper 18 June 2026. Search excerpt.
* [Ofgem and the Energy Ombudsman on AI](https://www.ofgem.gov.uk/blog/understanding-ai-energy-sector-benefits-risks-and-your-rights), 5 February 2026.
* [Government guidance on integrating AI in operational technology](https://www.nsa.gov/Press-Room/Press-Releases-Statements/Press-Release-View/Article/4347041/nsa-cisa-and-others-release-guidance-on-integrating-ai-in-operational-technology/), December 2025.
* [New York PSC inquiry into AI at utilities](https://dps.ny.gov/news/psc-launches-inquiry-use-artificial-intelligence-new-yorks-utilities), 17 September 2026.
* The Red Energy penalty of February 2025 was read on an AEMC page as a search excerpt.
**Deployments.**
* [Tapestry on Vector](https://x.company/blog/posts/tapestry-vector/), 20 August 2025 \[vendor].
* [iTnews on Essential Energy](http://www.itnews.com.au/news/essential-energy-turns-to-ai-for-safety-inspections-625091), 16 April 2026.
* [Watercare sensor rollout](https://www.watercare.co.nz/home/about-us/latest-news-and-media/watercare-s-wastewater-network-smartens-up-with-rollout-of-5000-sensors), 18 August 2025.
* [Trenchless Australasia on Sydney Water](https://www.trenchless-australasia.com/2023/02/22/sydney-water-takes-on-more-tech/), 22 February 2023. Search excerpt.
* [IBM case study on AGL](https://www.ibm.com/case-studies/agl-energy), undated \[vendor].
* [PolyAI on PG\&E](https://poly.ai/customers/pge), 2026 \[vendor].
* [Pulse2 on Kraken and Sierra](https://pulse2.com/kraken-launches-autonomous-agents-with-sierra-to-transform-utility-customer-service/), 12 June 2026 \[vendor], and [Sierra on Kraken](https://sierra.ai/customers/kraken), 2026 \[vendor].
* [CX Today on Kraken's agents](https://cxtoday.com/kraken-autonomous-agents), 12 August 2026. Independent trade analysis. It says anything that changes a balance, payment, tariff or debt should be a hard stop.
* [Kraken case study on Origin](https://www.kraken.tech/case-studies/origin), undated \[vendor].
* [Origin half-year 2026 presentation](https://www.originenergy.com.au/wp-content/uploads/65/Origin-Energy-HY26-Investor-Presentation.pdf), 12 February 2026. Search excerpt.
* [Kraken and Snowy Hydro partnership](https://www.kraken.tech/press-releases/snowy-secures-kraken-partnership), 15 July 2026. More than 1.6 million customers. No mention of AI agents.
* [Hansen AI Virtual Agent](https://www.financialcontent.com/article/gnwcq-2025-2-18-hansen-launches-ai-powered-virtual-agent-to-enhance-customer-experience), 18 February 2025 \[vendor]. Hansen's FY26 results of 19 August 2026 name City of Kingsport (US) as a deployment.
* [Salesforce Agentforce for Energy and Utilities](https://salesforce.com/agentforce/use-cases/energy-utilities/customer-service), 2026 \[vendor]. Pre-built subagents. Salesforce warns features may change.
* [Kurrant on Ofwat's AI plan](https://kurrant.com/kurrantly-news/ofwat-backs-ai-for-the-water-sector-as-data-centre-demand-strains-supply/), 18 June 2026.
**Gaps.** No volumes or cycle times were found for Australian billing exceptions, water alarms or contact centre journeys. Esri's agent products from the July 2026 conference need Esri's own announcement before they are named.