Frank Carnevale, Country Head, Canada for iGreenTree.

Electricity affordability isn't a slogan; it's the scorecard. As distribution companies absorb electric vehicle (EV) load growth, aging assets, extreme weather and a decarbonizing grid, every operational choice shows up on the bill. The practical path forward is deploying AI where it trims operating expenditures, defers capital expenditures and improves reliability without waiting for the next metering generation.

Starting With The Data You Already Have

Here's the punchline most miss: The vast majority of high value analytics can be proved on first-generation Advanced Metering Infrastructure (AMI 1.0) data. Existing interval reads and basic meter events—joined with an outage management system (OMS), a customer information system (CIS), supervisory control and data acquisition (SCADA) and a geographic information system (GIS)—are enough to unlock real savings.

AMI 2.0 (faster intervals, richer edge telemetry) is welcome, but it's not a prerequisite for value or affordability. Many utility executives believe that with AMI 2.0 comes great data maturity. It does not.

What Moves The Needle (Now)

1. Fault detection and predictive maintenance can result in fewer emergency truck rolls and longer asset life, including customers' assets. Learn normal loading and temperature signatures from AMI 1.0 and weather, surface deviations early, push work orders to a work management system (WMS) and share insights with customers.
2. Load forecasting and demand flexibility allow for smaller reserves and fewer surprise peaks. Blend weather, AMI 1.0 intervals, EV clustering cues and distributed energy resources (DER) patterns to time shift EV charging, pre cool buildings and pre commit demand response (DR) capacity.
3. Voltage optimization (CVR/VVO) can allow for a system wide kWh reduction without compromising power quality (PQ). Continuously tune feeder setpoints and verify savings in AMI 1.0.
4. Non technical loss analytics can result in lower leakage with minimal friction. Incorporate anomaly detection on AMI/CIS to find theft, crossed meters and chronic exceptions and close the loop with respectful field workflows.
5. Vegetation and asset risk targeting allow for fewer outages and overtime events. Combine outage history, asset condition and simple geospatial risk factors to prioritize the spans that matter.

AMI 1.0 Vs. 2.0: Don't Let 'Perfect' Delay Payback

Most affordability driving use cases do not require AMI 2.0 to begin. Examples that run well on 1.0 datasets include transformer loading and lifetime (intervals and temperature indices), feeder CVR verification (kWh and PQ compliance), EV clustering detection (evening ramp signatures), theft/exception analytics (meter events and CIS joins) and hosting capacity triage (load shapes and GIS topology). If you plan to wait for AMI 2.0, affordability is already paying the price.

AI On Both Sides Of The Meter

Treat the meter as a handshake, not a wall. It's always been amazing to me to see utilities not fully understand their customers' needs and customers not having a clue what utilities must do to keep the lights on.

• Utility Side: Predictive maintenance, CVR/VVO (volt/VAR optimization), loss analytics, fault location, isolation and service restoration (FLISR) and predictive outage, hosting capacity, and non-wires alternatives (NWA) screening.
• Customer Side: Feeder aware EV charging, premise level insights that drive time of use response, battery/thermal schedules that support the grid and lower bills, targeted outreach that reduces cost to serve.

When both sides are optimized, the system needs fewer emergency fixes and defers capital, and customers see tangible benefits. That's affordability.

From Data To Decisions

Utilities are data rich and outcome poor. AMI, SCADA, OMS, CIS, GIS—terabytes of data arrive daily, but value stalls in dashboards. The fix isn't another pilot; it’s an operating model.

• Productize use cases - owners, SLAs, release cadence and measurable KPIs.
• Adopt mature smart grid data software that turns AMI into operations (outages, transformer loading, hosting capacity, feeder stress) without ripping and replacing hardware.
• Finish the last mile - pipe recommendations into OMS/WMS/CIS where dispatchers, planners and customer teams decide hourly.

What I'm Hearing Across Canada

• EV charging management is table stakes. Unmanaged charging quietly overloads neighborhood feeders and compresses transformer lifetimes. Align charging windows to feeder capacity and rate design, keep it opt in and "set and forget." ROI shows up as deferred upgrades and fewer emergencies.
• Analytics must be prescriptive. Dashboards don't move trucks. Operators want ranked actions—which transformer first, which feeder gets CVR now, which customers to enroll in a new rate.
• Legacy CIS undermines affordability. Outdated stacks drive manual work, exceptions and poor integration with AMI/DER. Modern, AI enabled CIS can reduce call volumes, speed move in/move out, improve credit/collections and support dynamic rates and DER credits. Less rework, fewer escalations, better customer experience.

DERMs: Fortify Beyond The Fence Line

An increasing share of grid "muscle" sits behind the meter—rooftop solar, batteries, heat pumps, EVs, smart devices. They aren't rate based or utility owned, but they can be orchestrated.

AI enabled distributed energy resource management systems (DERMs) unlock four affordability levers—peak shaving via behind the meter batteries, feeder aware EV coordination, solar plus storage for resilience/microgrids and NWA screening to stack voltage support, capacity and reliability before pouring concrete.

The Affordability Shortlist

If the mandate is "keep rates flat where possible," fund this stack first:

• CVR/VVO with continuous learning
• Transformer loading and lifetime models tuned to EV clustering/electrification
• FLISR and predictive outage to pre position crews/parts
• Advanced loss and exception analytics
• EV managed charging with opt in incentives and feeder constraints
• Modern CIS with AI assistance
• Automated hosting capacity and NWA screening to target capital where it truly earns its keep

Quick Self Check

Ask, candidly:

1. Which feeders have verified CVR savings in the last 90 days?
2. How many truck rolls were avoided last quarter from AMI driven predictive alerts?
3. Where are EV clusters stressing transformers—and what actions did we take?
4. What's our 12 month loss recovery trend?
5. Which CIS exceptions fell because AI caught them upstream?
6. How many customer side assets are enrolled in feeder aware programs?

If these aren't on hand, you're sitting on affordability you haven't claimed yet. How mature is your utility with its data today?

Bottom Line

Before funding the next big build—or the next metering generation—run the math on the algorithms. Your cheapest capacity, quickest reliability lift and cleanest path to protect affordability may already be sitting in AMI 1.0 data—waiting to be orchestrated on both sides of the meter.

About The Author

Frank Carnevale, Country Head, Canada for iGreenTree.ai

Frank Carnevale is a senior executive with over 30 years of experience in cleantech, proptech, utilities, and energy. As Country Head, Canada at iGreenTree.ai, he leads national efforts to advance AI adoption and digital integration across utilities, power companies, and municipalities.

He has previously served as CEO of BHC Canada, CEO of Cleantech Power Corp., and Chief Growth Officer at Universal PropTech Inc., driving the convergence of clean technologies, building systems, and digital infrastructure. Frank also supports operational innovation as Innovations Lead at Dexterra On-Demand.

Throughout his career, he has led and originated more than $3 billion in energy-related transactions across regulated utilities, developers, and public-sector clients. He also serves on the Board of Governors at Ontario Tech University and continues to contribute to Canada’s energy and cleantech landscape.

Originally Posted on Forbes