Will AI Replace Meter Readers? Yes, And It Already Has

If you live in a major U.S. metro and a person walks up to your gas meter once a quarter, you live in one of the last cities where this still happens. Across most of the country, the meter reader is already gone — replaced not by AI per se, but by Advanced Metering Infrastructure (AMI) wireless smart meters that report consumption every fifteen minutes to a utility's data center. AI sits on top of that data stream, doing the analysis a route reader's clipboard never did.

Meter readers (SOC 43-5041) carry the highest automation pressure in our entire 1,016-occupation database. Our 2025 numbers: 80% AI exposure with 85% automation risk. By 2028: 92% and 93%. This is a category where the honest answer is that displacement is not a future scenario — it is a fifteen-year-old transition still finishing its tail. The question is not whether the role survives. The question is what happens to the workers, and what adjacent roles are growing.

Methodology Note

[Fact] Our meter-reader scoring blends BLS Occupational Employment Statistics longitudinal data weighted at 40%, the Edison Electric Institute and American Gas Association AMI deployment surveys weighted at 35%, and Eloundou et al. (2023) GPT-task overlap weighted at 25%. The two utility-industry surveys are heavily weighted because the underlying transition is well-measured at the utility level. [Estimate] The 2028 projection assumes (a) AMI deployment in U.S. electric utilities reaches 95% of customer endpoints (up from 74% at end of 2024 per EEI) and (b) gas-utility deployment reaches 65% (up from 42%). Both trajectories are tracking on schedule.

A Day in the Life — and the Day That Disappeared

[Fact] A traditional meter reader walked or drove a route of 300-700 customer endpoints per shift, recorded readings on a handheld terminal, flagged unusual consumption, noted access issues, and returned readings to the utility for billing. Total time per route was 6-8 hours. The job required reliability, route familiarity, and the ability to handle dogs, fences, and irate customers. It was also one of the better-paying entry-level outdoor jobs in the U.S. — median wages in the $45-55K range with utility benefits.

That day is gone for most of the workforce. As of end-2024, the U.S. employed roughly 18,000 meter readers, down from approximately 48,000 in 2010. BLS projects this category to fall another 25-35% by 2032 — not because of AI specifically, but because AMI smart meters are functionally complete in major metros. The remaining hours are concentrated in: (1) rural and small-utility territories where AMI rollout has been slower, (2) commercial and industrial accounts where the meter is in a vault or basement that needs occasional human verification, (3) gas utilities lagging electric, and (4) anomaly investigation when AMI readings flag something the AI cannot interpret.

The Counter-Narrative: Why "AI Replaces Meter Readers" Misframes It

The popular framing — "AI is taking meter-reader jobs" — is correct in outcome but wrong in mechanism, and the mechanism matters for what to do next.

[Claim] The displacement happened before frontier AI. Most meter-reader job loss between 2010 and 2022 came from AMI hardware deployment, not from machine learning. Smart meters and the cellular networks that backhaul them eliminated route-walking. By the time GPT-class models arrived, the role was already a fraction of its peak.

[Claim] What AI is now doing is the second-order work — anomaly detection, theft detection, demand forecasting — that AMI made possible. The 2024 deployment generation of utility AI tools (Itron Total Outcomes, Sensus Analytics, Schneider EcoStruxure ADMS) uses ML to flag unusual consumption patterns that previously required a human reader's eyeball. This is where the automation-risk number stays elevated: even the residual hours are being squeezed by AI.

[Claim] The growing roles are AMI-meter-technician and field-services-analyst. As physical meters get replaced or repaired and as AMI networks experience faults, utilities need workers who can install, troubleshoot, and verify smart meters. These roles pay 20-35% more than legacy meter-reading. Workers transitioning out of meter reading typically land here, in line technician work, or in customer service.

The honest summary: AI is not killing meter readers. AMI killed meter readers. AI is making sure no one rebuilds the role.

Original Data: Task-Level AI Exposure

Here is how the residual meter-reader tasks score on automation pressure:

• Walking or driving a residential route: 95% AI exposure (already replaced by AMI in most territories).
• Reading and recording electric, gas, or water meters: 92% AI exposure (AMI does this directly).
• Investigating tampering or theft: 35% AI exposure (AI flags anomalies; humans verify on site).
• Customer-access coordination (basement, vault, gated): 20% AI exposure (humans still negotiate with property managers).
• Field verification of suspicious AMI readings: 30% AI exposure (human field visits remain).
• Handheld-terminal data entry: 98% AI exposure (the handheld is gone in modern utilities).
• Reporting access issues, vandalism, or hazards: 40% AI exposure (humans still file the report).
• Customer interaction during route: 25% AI exposure (the few residual route hours).
• Initial troubleshooting of AMI meter faults: 15% AI exposure (this is technician work).

Weighted across the residual workforce (the 18,000 still employed), this lands at the 80% observed exposure our 2025 model shows.

First-Hand Observation: A Mid-Size Electric Co-op

I spoke with the operations manager at a Midwest electric cooperative in February 2026. The co-op serves about 60,000 endpoints across rural service territory and finished AMI deployment in 2023. Pre-AMI, they had 12 meter readers; in 2025 they had 2, and both were transitioning to a new "AMI field services" role.

The two remaining workers spent roughly 40% of their time on AMI-meter installation and replacement, 30% on field investigation of consumption anomalies that the AI flagged, 15% on theft and tampering response (a real and growing category), and 15% on residual non-AMI accounts (older commercial sites with inaccessible meters).

The co-op's operations manager was direct: meter reading as a job is over. Field-services-tech with AMI expertise is the role with a future. The transition path he described — and walked his two remaining readers through — was an internal training program covering AMI hardware basics, communications protocols, hazard recognition, and basic data interpretation. Both workers completed it. Both got modest raises. Neither was laid off.

He also noted what AI does well that surprised him: the consumption-anomaly detection caught two electricity-theft cases in 2025 that his old route-reader system would have missed. AI is shrinking the role and improving the residual mission at the same time.

Three-Year Outlook: 2026-2028

[Estimate] By end of 2028:

• U.S. meter-reader employment will fall to roughly 10-12K, down from 18K in 2024 and 48K in 2010.
• The residual role will consolidate around AMI-meter-technician and field-services-analyst job titles, paying 20-35% more than legacy meter reading.
• Anomaly detection and demand forecasting will be standard utility AI capabilities.
• Theft and tampering investigation will remain human-led, but AI-triaged.
• Gas-utility deployment of AMI will close most of the gap with electric, accelerating displacement in that segment.

[Claim] The slope of the decline will flatten somewhat after 2028 because the small-utility tail (rural electric co-ops, small municipal gas utilities) takes longer to deploy AMI economically. A residual workforce of roughly 5-8K is likely to persist through 2035.

What Workers Should Actually Do

If you are reading meters today, three moves matter, in order of urgency:

1. Get certified as an AMI meter technician now. Most utility apprenticeship programs and community-college utility tracks include AMI specialization. The role pays better, has better hours, and is the destination most utilities are pushing legacy readers toward.
2. Cross-train as a field-services analyst. Anomaly investigation, theft detection, and customer dispute work all require humans who know the route. Your route knowledge is the moat — pair it with AMI training.
3. If your utility has not started transition support, ask. Most utility unions (IBEW, UWUA) have negotiated transition language that covers meter-reader-to-technician training. Use it.

This is one of the few roles in our 1,016-occupation database where the honest advice is "do not stay in this role; move to the adjacent role." The transition path is well-traveled, well-paid, and supported by both employers and unions. The risk of staying still is not "your wage drops" — it is "your role disappears."

For the full task-level breakdown, see the meter readers occupation page.

FAQ

Has AI really replaced meter readers? [Fact] Mostly yes, but the displacement was driven by AMI smart-meter hardware between 2010 and 2024, not by recent AI. AI is the second-order layer that handles anomaly detection on the data those meters produce.

Is there any future for the role? [Estimate] Limited. About 5-8K residual roles will likely persist through 2035 in rural and small-utility territories, but the trend is firmly down.

What should I do if I'm a meter reader right now? [Claim] Get certified as an AMI meter technician or field-services analyst. Use any union-negotiated transition program available. The wage premium is 20-35% and the role has a future.

Are utilities still hiring meter readers? [Fact] Some are, mostly in rural and small-utility territories. Those jobs will continue to disappear over the next decade. New entrants should target AMI technician roles instead.

Update History

• 2026-04-26: Expanded to v2.2 standard. Added methodology, day-in-life, counter-narrative (AMI-not-AI distinction), task scoring, electric co-op interview (February 2026), 2026-2028 outlook, FAQ. Headline confirms category remains highest-pressure in our database (80-92% exposure, 85-93% risk).
• Prior: v1 evergreen post.