
Advanced metering infrastructure (AMI) is usually justified in operational terms. A Water Finance and Management piece by WFM Staff on AMI argues that the conversation is moving from data collection to data usage. Its summary lists the benefits most utilities already expect, including more accurate billing, leaks identified sooner, and the fact that "utility staff no longer have to walk meter routes." Those are operational gains. How large each one is for a given utility depends on its own baseline, and the summary does not supply one.
Collection Approval Is Not Use Approval
Most AMI business cases are approved against a bundle of purposes. The board authorizes the capital and accepts a set of operating claims, and the approval covers the act of collecting interval reads. It rarely specifies every downstream use. Once a register of frequent consumption readings exists, requests tend to follow. Customer service wants answers for high-bill calls, finance wants demand models, a leak program wants account-level flags, and an enforcement group may want to identify irregular use. Each request can be reasonable on its face. Each one is a new purpose applied to the same records.
Why Secondary Use Changes the Risk
The same read can carry different reliability requirements depending on the job. A billing read has to be accurate within the register's rated limits and correctly timestamped. A demand model may tolerate gaps if the gaps are flagged. A flag that affects an individual household needs a documented error margin and a way for the customer to contest the result. When data moves from billing into a decision about a person, an error that was tolerable at the billing stage becomes a consequence.
Conventions set in the billing system also travel with the data. Timestamp rules, read frequency, and register units were chosen for invoicing, and an analyst building on them may not know those choices exist. A consumption spike can reflect a change in read interval rather than a change in use. The gap between what the meter measured and what the analysis assumes is where most secondary-use errors sit, and it is rarely visible in the output.
Write the Approval Record Per Use
A practical control is a short use register. For each approved use, the register can list the data fields drawn on, the minimum accuracy assumed, the groups allowed to see results, the conditions under which a result may be acted on, and a review date. When a new request arrives, the first question is whether it fits an approved use or needs a new line. The aim is not to require a fresh board vote for every query. It is to require a documented decision for every new purpose.
The register also helps staff separate a question about the data from a question about a person. A request to understand system-wide demand can usually proceed under an existing line. A request to identify a single account for follow-up needs its own line, with its own error handling. Whether a given use requires customer notice or other disclosure depends on the state and local rules that apply to the utility. Those rules fall outside this article, and the utility's counsel should answer them.
What the Benefit Claims Do Not Settle
The benefits in the summary are measured against something. Staff time is compared with the walk-route effort that existed before AMI. Billing accuracy is compared with the error pattern of the earlier read method. Leak detection speed is compared with the time it took to find leaks before the system was installed. A business case that cites these gains without the earlier numbers is describing a direction, not a result. Where the pre-deployment record is thin, the claim should be labeled as an estimate, and the utility should say what evidence would change it.
None of this argues against AMI. It argues that the data's second and third uses deserve the same written scrutiny as the first. A utility that treats each new use as a fresh decision, with its own accuracy assumption and its own access rule, will be able to answer later questions about why a number was used and for what. A utility that treats the meter fleet as a single approved asset will find that its records are doing jobs nobody wrote down.