Smart meter rollouts are running into a workforce ceiling, not a technology one.

Detection algorithms already flag theft and faults with near-perfect accuracy. The constraint utilities are discovering is the number of trained people available to act on what the meters tell them.

Nigeria's Power Force launched this month with a simple premise: 5,000 young Nigerians trained to install and maintain smart meters, starting in Abuja before spreading across all six geopolitical zones. Applications for the first cohort opened on 4 July and closed two weeks later. It is a jobs programme and a metering programme in the same breath, and that pairing is the point.

Meters do not install themselves

Every headline target assumes a workforce that can execute it. Romania's plan to add over 800,000 smart meters needs crews on the ground. West Bengal's rollout, covering close to two crore consumers, needs installers who can move from government campuses to large consumers to households in sequence. Ofgem's direction requiring suppliers to complete rollout by the end of 2030 and replace ageing communications hubs assumes a technician base large enough to do it on schedule, not just a submitted deployment plan.

The data is already ahead of the workforce

Maharashtra's state utility MSEDCL detected 19,599 electricity theft cases over four months, with 11,032 of them found in July alone. The estimated value of that theft was put at Rs 30.36 crore. Recovery so far stands at Rs 1.85 crore, compounded from 4,959 cases. That gap between cases found and cases resolved is not a detection failure. It is a processing one: every flagged case still needs a person to visit the site, confirm the tamper, and close it out.

Academic work backs up the point that detection itself is a solved problem in most respects. Machine learning approaches using electricity, gas and water consumption data have reached accuracy above 95 percent with area under the curve around 0.99 using tree-based ensemble methods. Earlier work using artificial neural networks reached similar conclusions analysing consumption patterns alone. The models are mature. What they produce is a list of names and addresses that someone has to physically go and check.

Training as infrastructure, not overhead

Power Force treats workforce capacity as a deliverable in its own right, alongside meters and comms hubs, rather than an assumption baked into a procurement timeline. That framing matters wherever a rollout is scaling faster than the pool of people qualified to install, inspect and investigate. A theft detection model that flags a thousand suspect accounts a month is only useful if there are enough field staff to work through them before the backlog makes the list meaningless. The same logic applies to hub replacement schedules and to large residential rollouts where installation quality determines whether the meter reports reliably for the next decade.

Power Force will train 5,000 young Nigerians to support the country's smart meter rollout, creating jobs while strengthening ongoing power sector reforms.

State House, Abuja

For utilities and governments planning the next phase of AMI investment, the lesson is not that analytics need improving. It is that the operational layer behind the analytics, the people who convert an alert into a resolved case or a delivery plan into an installed meter, needs the same deliberate investment as the technology itself.