InfrastructureIssue 073 records
Utility Management
A thin day. Oshawa Power, a municipally owned Ontario utility, extended its AI assistant to move requests and balance checks behind multi-factor verification. By contrast, Plumas-Sierra Rural Electric Cooperative told members it will not use AI for customer service. Brownsville PUB in Texas began AMI-based leak alerts, a rule-based base layer for future metering analytics. No item reports measured results; US gas distribution, stormwater AI and OT/AI security had no fresh evidence.
- Evidence records
- 3
- Cross-source patterns
- 1
- Evidence classes
- 2 independent reporting1 public-sector association guidance
- Outcomes
- 2 emerging1 cautionary
- Source freshness
- 3 new this fortnight
- Research completed
- 2026-11-05
Choose a role to see its takeaway beside every record in the ledger.
Patterns across the evidence
Synthesis · Lighthouse Advisory interpretation
Utilities are drawing explicit, public lines on member-facing AI, in opposite directions
Oshawa Power extended its AI assistant to account transactions behind multi-factor verification, while Plumas-Sierra REC told members it will not use AI for customer service. Both keep humans reachable. Neither reports measured outcomes, so the choice is driven by member expectations and controls rather than evidence of results.
Operating questionWhich account interactions will members accept from AI, and what verification and human-access guarantees are published?
Supporting evidenceOshawa Power (municipally owned local distribution company)Plumas-Sierra Rural Electric Cooperative (PSREC)
Evidence ledger
Full record · every source keeps its link and limitations
Plumas-Sierra Rural Electric Cooperative says it will not use AI for customer service
In its November 2026 manager's message, PSREC says it and peer utilities are watching AI uses such as reviewing aerial photographs for line maintenance, but states it is not interested in using AI for customer service, because members are best served when they can quickly reach a human, day or after hours.
Why it matters, evidence and limitations
- Why it matters
- A member-owned co-op publicly ruling out AI in customer service is a direct counterpoint to municipal and co-op chatbot and voice-agent rollouts. It shows member expectations and governance can limit AI scope regardless of technical readiness.
- Evidence and measured results
- Policy statement by the utility's management in a member newsletter. No data, survey or cost analysis is offered. The day of publication within November is not stated; 1 Nov is used as the month start.
- Limitations and uncertainty
- Single small co-op; position statement, not evaluation. No details on what AI uses it is pursuing.
Oshawa Power upgrades AI virtual assistant Watt to handle moves, balances and bill due dates
Oshawa Power launched an upgraded virtual assistant, Watt, that uses AI to understand questions and complete tasks 24/7, including starting or stopping service for moves and checking balances or bill due dates. The utility says move requests and balance inquiries are more than 25% of customer service calls, and Watt uses multi-factor verification before account updates or sharing account information. Human agents remain available.
Why it matters, evidence and limitations
- Why it matters
- A municipally owned utility moving AI from FAQ answers to account transactions is directly transferable to US municipal electric and water billing offices, though Ontario's regulatory and privacy context differs.
- Evidence and measured results
- Local news report based on the utility's announcement. The 25% figure is the utility's share of call volume for these request types, not a measured deflection result. No accuracy, containment or satisfaction data. The utility's own pages returned HTTP 403 to our fetch.
- Limitations and uncertainty
- No results yet; vendor and model not named; utility site not accessible to verify directly.
Brownsville PUB starts AMI-based customer leak alerts by text and email
On 1 Nov 2026 BPUB began texting or emailing customers when AMI data shows prolonged water use, with notices every 24 hours, weekly or biweekly depending on consumption. The AMI program, funded partly by a $5 million 2023 WaterSMART grant, replaces over 17,500 manual meters and retrofits almost 40,000 more, with full deployment expected by 2027 and a target of conserving 2,103 acre-feet a year. Alerts come only from short code 96681 and never request payment.
Why it matters, evidence and limitations
- Why it matters
- A US municipal utility turning AMI interval data into customer-facing alerts is the data foundation that metering and billing AI depends on. The alert logic described is a continuous-use rule, not AI; relevance to AI is as a prerequisite and as an anti-scam design lesson.
- Evidence and measured results
- Trade article citing KRGV, E Source and APPA. Grant, meter counts and acre-feet target are planning figures; no alert results yet. Detection method described as prolonged-use triggers, not machine learning.
- Limitations and uncertainty
- Not an AI deployment; no measured leak or water savings yet; secondary source.
How to read this edition
Source findings, measured results and limitations come from the cited publications. Patterns, operating questions, role takeaways and implementation considerations are Lighthouse Advisory interpretation, stated as questions to validate locally rather than guaranteed outcomes. Vendor and operator claims are labeled as claims. Full research method.
- Independent reporting
- Independent reporting with attributable sources but without a formal evaluation design.
- Public-sector association guidance
- Practitioner guidance or an association-supplied case; not independent outcome evidence.