# What Is Automatic Meter Reading (AMR) and How Does It Work? Networks, Costs and Pilot Guide

> Automatic meter reading (AMR) collects water, electricity and gas readings without site visits. Learn how it works, LoRaWAN vs NB-IoT and a 6-step pilot.

Source: https://digitalbridge.com.tr/en/blog/automatic-meter-reading-guide.html
Publisher: Digital Bridge Teknoloji (https://digitalbridge.com.tr)
Topic: IoT & Hardware
Published: 2026-09-24 · Updated: 2026-10-22

Automatic meter reading ([AMR, known in Türkiye as OSOS](https://digitalbridge.com.tr/en/glossary.html#osos)) means fitting water, electricity, gas and heat meters with a communication device that reads the index at set intervals and sends it to central software. Nobody walks door to door, billing errors and disputes fall, and a stopped meter or an illegal connection shows up long before the next billing cycle.

## Key takeaways

- Automatic meter reading (AMR) moves water, electricity, gas and heat meter readings to central software at set intervals, with no site visits.
- Every AMR system has four layers: the meter, the communication device, the network and central software linked to billing.
- Network options are GPRS/4G, NB-IoT, LoRaWAN and wired M-Bus/RS-485; the right choice is made with on-site signal measurements.
- Türkiye's average drinking water loss rate was 31.6% in 2024; metropolitan municipalities must reach at most 25% by 2028.
- A successful pilot runs in parallel with manual reading for at least one full billing cycle, and scales only once every discrepancy is explained.

Getting automatic meter reading right depends on planning three things together: the meter types, the network coverage on site and the billing system the data must reach.

## Why does manual meter reading fall short?

In a water utility, an organised industrial zone or a multi-block residential estate, manual meter reading repeats the same cycle every month. A crew drives out, hunts for the key to the meter cabinet, and estimates the reading wherever a door stays shut. Every link in the manual reading chain, from the handheld terminal to the billing system, can fail: digits are misread, entered against the wrong account or never captured.

The deeper problem with manual reading is timing: if a meter stops, a pipe bursts or someone bypasses the meter, a monthly read only reveals it weeks later. The consumption in between is either never billed or disputed by the customer. And with one data point per month, management cannot analyse night-time flow, zone-level losses or consumption patterns.

The same infrastructure can monitor pumps and reservoirs as well as meters; we cover the pump side in [remote pump station monitoring](https://digitalbridge.com.tr/en/blog/remote-pump-station-monitoring.html) and the reservoir side in [water tank level monitoring](https://digitalbridge.com.tr/en/blog/water-tank-level-monitoring.html).

Automatic meter reading changes the manual-reading picture in two ways. It cuts the cost and error rate of reading, and, more importantly, it raises the reading frequency so that problems surface when they happen rather than at the end of the month.

## What does it cost not to switch to automatic meter reading?

On the water side, the cost of manual reading is measurable because water supplied but not billed, known as [non-revenue water](https://digitalbridge.com.tr/en/glossary.html#non-revenue-water), is recorded:

> According to Türkiye's Ministry of Agriculture and Forestry, the average water loss rate in drinking water networks fell from 39% in 2015 to 31.6% in 2024, and metropolitan and provincial municipalities must bring it down to at most 25% by 2028. ([Ministry of Agriculture and Forestry — Urban Water Efficiency](https://www.suverimliligi.gov.tr/home/KentselSuVerimliligi))

The water loss target is a legal obligation, not a recommendation. Türkiye's 2014 [Regulation on the Control of Water Losses in Drinking Water Systems](https://www.tarimorman.gov.tr/Duyuru/1461/Icme-Suyu-Sistemlerinde-Su-Kayiplarinin-Azaltilmasina-Yonelik-Is-Termin-Plani-Genelgesi-Yururluge-Girmistir) sets limits of 30% by 2028 and 25% by 2033 for smaller municipalities too. Commercial losses such as misreads, under-registering meters and theft stay largely invisible without accurate, frequent meter data.

Globally the picture is starker still. The [World Bank's 2018 assessment of non-revenue water](https://blogs.worldbank.org/en/ppps/what-do-private-companies-look-performance-based-non-revenue-water-project) puts water produced but never billed at 126 billion cubic metres a year, equivalent to nearly $40 billion in annual losses from waste and foregone revenue.

In energy, measurement is the precondition for efficiency. According to the [Turkish Ministry of Energy's Second National Energy Efficiency Action Plan](https://enerji.gov.tr/Media/Dizin/BHIM/tr/Duyurular/T%C3%BCrkiyeninEnerjiVerimlili%C4%9Fi2030StratejisiveIIUlusalEnerjiVerimlili%C4%9FiEylemPlan%C4%B1_202401161407.pdf) (2024), buildings and services accounted for 32.6% of Türkiye's final energy consumption in 2022, and industry for 31.6%. Electricity distribution losses are significant too: according to [TEDAŞ's 2024 Turkish Electricity Distribution Sector Report](https://www.tedas.gov.tr/FileUpload/MediaFolder/ff204253-90d1-4560-8946-cdc529dceff2.pdf), Türkiye's combined technical and non-technical loss rate was 11.02% in 2023.

For property managers, industrial zones and multi-building campuses, building and industrial consumption cannot come down without seeing it per flat, building or line. For industrial customers, reactive consumption read frequently from the electricity meter reveals the risk of a [reactive power penalty](https://digitalbridge.com.tr/en/blog/reactive-power-penalty.html) before the bill arrives.

## How does automatic meter reading (AMR) work?

Every automatic meter reading system has four layers, and each one needs its own decision:

1. **The meter.** Mechanical or ultrasonic water meters, single- or three-phase electricity meters, gas meters or heat meters. Whether the meter has a pulse output, an optical port or an M-Bus/RS-485 interface decides which reading device can be used.
2. **The communication device.** A module fitted to or beside the meter reads the index and transmits it. At this layer you decide between battery and mains power (Digital Bridge's battery-powered models target a 5–10 year battery life, depending on reading frequency), and whether the device should detect cover removal and tampering. For damp spots such as meter cabinets and chambers, the enclosure's [IP67 rating](https://digitalbridge.com.tr/en/blog/ip-rating-explained.html) is chosen here too.
3. **The network.** Data reaches the centre over a cellular network (GPRS/4G or NB-IoT, narrowband IoT), a private network the organisation runs on its own [LoRaWAN gateways](https://digitalbridge.com.tr/en/blog/lorawan-gateway-guide.html), or a wired M-Bus/RS-485 line.
4. **The central software.** It validates readings, flags missing ones, raises alerts for zero consumption or unusual spikes (using [anomaly detection](https://digitalbridge.com.tr/en/blog/anomaly-detection-guide.html)), and passes billing data to the billing or customer system at the end of each period.

The layer organisations most often neglect in AMR projects is the fourth: the central software and its link to billing. Automating reading while someone still copies data into a spreadsheet leaves half the cost in place. Billing integration should be designed from day one, which is why a [system integration](https://digitalbridge.com.tr/en/technology-hardware/system-integrations.html) review of how your customer system accepts data belongs at the start of the project.

## Which communication technology should you choose for AMR?

There is no single "best" network for automatic meter reading; the right choice depends on coverage, meter density and who will operate the infrastructure.

| Technology | Best suited to | Watch out for |
|---|---|---|
| GPRS / 4G | Scattered customers, existing operator coverage | Per-SIM subscription; high power draw on battery devices |
| NB-IoT | Basements, meter cabinets and other hard-to-reach points; long battery operation | Operator NB-IoT coverage must be confirmed locally |
| LoRaWAN | Dense meter areas where the organisation runs its own network | Gateway placement and maintenance fall to the organisation |
| M-Bus / RS-485 (wired) | Many meters in one building, estates and industrial zones | Requires cabling; far easier in new builds |

Cellular IoT is no longer experimental. The [Ericsson Mobility Report (June 2026)](https://www.ericsson.com/en/reports-and-papers/mobility-report/dataforecasts/iot-connections-outlook) says cellular IoT connections reached around 4.5 billion at the end of 2025, and that NB-IoT and Cat-M continue to be rolled out for low-cost devices with long battery lives. We compare power, coverage and operating models in more depth in our [LoRa vs NB-IoT vs cellular guide](https://digitalbridge.com.tr/en/blog/lora-vs-nb-iot-vs-cellular.html).

The practical rule for choosing a network: decide with field measurements, not at a desk. Skip measuring signal at the five hardest meter locations in the pilot area (a basement, a metal cabinet, an underground chamber) and you risk the most expensive mistake at roll-out.

## How do you run an AMR pilot? A six-step framework

1. **Define the goal in numbers.** Cutting reading costs, reducing commercial losses and monitoring night-time consumption each need different reading intervals and reports. Write down how success will be measured before you start.
2. **Build a meter inventory.** Make, model, size, age and interface are needed before any device can be chosen. Replace faulty meters first; reading a bad meter remotely just delivers bad data faster.
3. **Choose a representative pilot area.** Not just an easy neighbourhood, but one that includes difficult coverage points. If you have a [geographic information system (GIS)](https://digitalbridge.com.tr/en/data-analytics/geographic-information-systems-gis.html) layer, placing meters on the map makes both the pilot and the roll-out easier; we explain how to set up that layer in our guide to [GIS for municipalities](https://digitalbridge.com.tr/en/blog/gis-for-municipalities.html).
4. **Set the reading interval to match the goal.** Daily reads are often enough for billing; loss analysis needs hourly profiles. Higher frequency costs battery life and airtime, so make that trade-off deliberately.
5. **Define alerts and owners.** Who receives zero-consumption, reverse-flow, tamper and spike alerts, and who acts on them within how many hours? A system that raises alerts nobody owns is soon ignored.
6. **Test billing transfer, then scale.** Run the pilot in parallel with manual reading for at least one full billing cycle. Do not scale up until every discrepancy is explained.

Guides to other metering, pump and sensor projects sit together in our [IoT & Hardware hub](https://digitalbridge.com.tr/en/blog/topic/iot-donanim.html). The water-loss side of the pilot framework, covering district metered areas and minimum night flow analysis, is covered in our [guide to non-revenue water loss analysis](https://digitalbridge.com.tr/en/blog/water-loss-analysis.html).

Hourly AMR data serves forward planning as well as billing. We explain how it becomes a site or regional load forecast in [electricity load forecasting](https://digitalbridge.com.tr/en/blog/electricity-load-forecasting.html).

## Common mistakes in automatic meter reading projects

- **Looking only at the device price.** SIM subscriptions, battery replacement, gateway maintenance and software operation make up a large share of total cost.
- **Forgetting personal data.** Household consumption data is personal data under KVKK, Türkiye's data protection law (broadly comparable to the GDPR). Decide early who can access it, how long it is kept and how customers are informed; setting this up through [data protection compliance consultancy](https://digitalbridge.com.tr/en/consultancy/data-protection-compliance.html) is easier than fixing it later.
- **Skipping tamper detection.** A reader that cannot detect cover removal or magnetic interference will struggle to reveal theft.
- **Leaving the field team out.** Meter readers do not disappear; they move to fault finding, maintenance and theft checks. If that shift is not planned, it creates resistance.

On the electricity side, the regulatory changes that directly affect industrial users (the 10 MWh threshold and the smart meter replacement timetable) are summarised in our article on the [automatic meter reading obligation in Turkey](https://digitalbridge.com.tr/en/blog/automatic-meter-reading-obligation-turkey.html).

Automatic meter reading is also the foundation of in-building energy management on spread-out campuses. We cover combining sub-meter data with the BMS across hospital blocks in our guide to [hospital building energy management](https://digitalbridge.com.tr/en/blog/hospital-building-energy-management.html).

Automatic meter reading is the measurement layer of a digital distribution network. We explain how it combines with GIS, SCADA and field service in our [guide to digital transformation in energy utilities](https://digitalbridge.com.tr/en/blog/energy-utility-digital-transformation.html).

## How does Digital Bridge deliver automatic meter reading?

At Digital Bridge, we treat automatic meter reading as a system designed for your site rather than a box off the shelf. Our electronics, enclosure design, embedded software and web/mobile panel come from one team (we describe the stages in our [IoT product development process](https://digitalbridge.com.tr/en/blog/iot-device-development-process.html) guide), so when something fails or needs changing, there is one party responsible.

- **We start with a free technical feasibility study.** We review your meter inventory, site coverage and billing system, and set out in writing which communication route makes sense.
- **We design and manufacture the communication device for your site.** Our [AMR meter reading system](https://digitalbridge.com.tr/en/industrial-products/osos-meter-reading.html) uses GPRS/4G, LoRaWAN, NB-IoT or wired M-Bus/RS-485 depending on conditions on the ground. For requirements beyond that, our [IoT device design and manufacturing](https://digitalbridge.com.tr/en/technology-hardware/device-manufacturing-iot.html) service takes over.
- **We deliver software and integration together.** The management software shows the latest reading, consumption charts, device health (signal, battery, last contact) and anomaly alerts in one panel, and passes billing data to your billing system via API. Pump stations, reservoirs and other sites can join the same [remote monitoring platform](https://digitalbridge.com.tr/en/blog/remote-monitoring-iot-platform.html).
- **We work in your sector's terms.** For water utilities and distribution companies we shape the project through our [energy and utilities solutions](https://digitalbridge.com.tr/en/industry-solutions/energy-utilities.html); for municipalities, through our [public sector and municipal solutions](https://digitalbridge.com.tr/en/industry-solutions/public-sector-municipalities.html).
- **We make on-site consumption visible too.** Factories and campuses that want to track electricity, water and gas per line use the same data foundation through our [energy monitoring and management](https://digitalbridge.com.tr/en/industry-40/energy-monitoring-management.html) service; we cover that side in our [factory energy monitoring](https://digitalbridge.com.tr/en/blog/factory-energy-monitoring.html) guide.

Digital Bridge AMR projects always follow the same order: needs analysis, a written proposal setting out scope, phases and cost, a pilot, then roll-out.

One of the most valuable uses of remotely read meter data is loss analysis; our article on [electricity theft detection](https://digitalbridge.com.tr/en/blog/electricity-theft-detection-analytics.html) shows how load profiles and meter event logs bring suspect customers to the surface.

For councils, automatic meter reading is a cornerstone of managing water loss and of a wider programme of [smart city applications](https://digitalbridge.com.tr/en/blog/smart-city-applications.html).

## How do you get started with automatic meter reading?

You do not need a big decision to get started with automatic meter reading. Prepare your meter inventory (types, quantities, rough locations) and the name of your current billing system, then get in touch through our [contact page](https://digitalbridge.com.tr/en/contact.html). We will plan signal measurements in a pilot area with you and find out, with real data, which technology will actually work on your site.

## Frequently Asked Questions

### Do we have to replace our existing meters for automatic meter reading?

Not always. Meters with a pulse output, optical port or M-Bus interface can take a reading module. For very old mechanical meters without any interface, replacing the meter is often the more economical and reliable option; that decision is made once the meter inventory is complete.

### What is the difference between AMR and AMI?

AMR is essentially one-way: readings travel from the meter to the centre. AMI (advanced metering infrastructure) adds two-way functions such as sending commands to the meter, for example closing a valve or updating a tariff. For many organisations, reliable one-way reading is the right starting point; two-way functions can be added once the need is clear.

### How often should meters be read?

It depends on the goal. For billing alone, daily or less frequent readings may be enough; loss analysis and night-flow monitoring need hourly consumption profiles. Reading frequency directly affects battery life and communication costs, so set it alongside the goal. During the pilot, different intervals can be tried to find the right balance from real data.

### Will the system work with our existing billing software?

It is designed to pass billing data to your existing billing system via API. At the start of the project we look at the format your customer system accepts; where there is no direct connection, we build an intermediate transfer layer. During the pilot, remote and manual readings are compared side by side, so every discrepancy in the transfer is spotted and fixed before billing depends on it.

### Is consumption data covered by data protection law?

Consumption data that can be linked to a household or person is personal data under KVKK, Türkiye's Personal Data Protection Law. Access rights, retention periods and customer notices should be planned at the outset and built into the system design. Because an hourly consumption profile can reveal a household's daily routine, access to detailed data should be limited by role.

### Should we choose LoRaWAN or NB-IoT for meter reading?

It depends on the site. LoRaWAN suits dense meter areas where the organisation wants to run its own network, although gateway placement and maintenance then fall to the organisation. NB-IoT stands out at hard-to-reach points such as basements and meter cabinets, provided operator coverage is confirmed. The most reliable decision comes from signal measurements at the hardest meter locations in the pilot area.

### How long does a battery-powered reading device last?

Battery life depends largely on how often the device reads and transmits. Digital Bridge's battery-powered AMR devices target a 5–10 year battery life depending on reading frequency, and battery level is shown in the management panel. Frequent reads such as hourly profiles drain the battery faster, so the reading interval should be set together with the project goal.