METER INTEGRITY · PREPAID & POSTPAID
A PRODUCT OF OT LIFESTYLE (PTY) LTD
MeterTruth pairs a secure element inside every meter with continuous edge analytics, so tampering, bypass and billing-integrity issues surface as verified events — not disputes. A policy-aware rules engine applies indigent and by-law exceptions automatically, before a case ever reaches a call centre.
7 pilot deployments scoped and ready to launch — pending onboarding support from OpenAI, Claude and Google Cloud Platform AI models.
National Treasury revenue-protection circulars and municipal consumer debt reports point to the same structural problem: prepaid bypass and postpaid meter tampering don't just cost individual units of electricity or water — they corrode the revenue forecasts municipalities depend on to plan infrastructure spend. Manual meter audits catch a fraction of cases, months after the tampering began, and often after the household has already accumulated debt it cannot realistically repay.
MeterTruth is built on published municipal finance and revenue-protection research, not a hunch. These are the reference points OT Lifestyle uses to size the problem, calibrate the tamper-detection model, and defend the case for a pilot to funders, municipal partners and AI model providers alike.
National Treasury's local government revenue-management circulars repeatedly flag electricity and water non-technical losses — meter bypass, tampering, and unbilled connections — as a structural drag on municipal cash flow that undermines infrastructure budgeting.
SOURCE: NATIONAL TREASURY MFMA CIRCULARS · REVENUE MANAGEMENT SERIES
Sector research into South African electricity distribution consistently identifies smart, tamper-evident metering — paired with automated analytics rather than manual audit cycles — as the highest-leverage lever municipalities have to reduce technical and non-technical losses.
SOURCE: SA UTILITY DISTRIBUTION-LOSS RESEARCH LITERATURE
Municipal indigent-policy frameworks require that subsidised consumption never be mistaken for tampering. MeterTruth's rules engine is designed directly around this requirement, checking every flagged event against registration data before a case is ever opened.
SOURCE: MUNICIPAL INDIGENT POLICY & BY-LAW FRAMEWORKS
Independent research on IoT-enabled infrastructure shows that combining edge-level anomaly detection with cloud-hosted large language models for case reasoning produces materially fewer false positives than either approach used alone — the core architecture MeterTruth applies to tamper detection.
SOURCE: APPLIED IOT / EDGE-AI INFRASTRUCTURE RESEARCH
Legal guidance on the Protection of Personal Information Act supports a model where the municipality remains the responsible party for consumption data, with the platform operator processing signed readings and tamper-event flags under a defined data-sharing agreement.
SOURCE: POPIA (ACT 4 OF 2013) GUIDANCE MATERIALS
The same tamper-resistant secure-element chips that anchor trust in banking cards and payment terminals are increasingly used in metering hardware to make a reading cryptographically provable, not just reported — the hardware foundation MeterTruth is designed around. "Banking-grade" here describes the security design principle and chip technology, not a banking certification held by MeterTruth.
SOURCE: SECURE-ELEMENT HARDWARE STANDARDS LITERATUREA tamper-resistant chip signs consumption data at source, so a reading can be cryptographically verified rather than taken on trust — the same integrity model used in banking hardware, applied to the meter.
The meter itself watches for magnet interference, cover tampering, and load-signature anomalies in real time, flagging events as they happen rather than at the next manual read.
Indigent registration status and municipal by-law exceptions are checked automatically before any case is escalated, so a subsidised household is never mistakenly flagged as a tamper case.
Confirmed events generate a documented evidence pack — timestamped, signed, and ready for a revenue-protection case — instead of a technician's disputed word against the resident's.
The secure element signs each consumption reading at the point of capture, creating a verifiable chain of custody for the data.
On-meter analytics continuously watch for tamper signatures — magnet interference, cover opening, load anomalies — without waiting for a manual visit.
Flagged events are checked against indigent status and by-law exceptions before escalation, keeping the case list fair and legally defensible.
Confirmed, policy-cleared cases generate an evidence pack and route to the revenue-protection team's dashboard, ranked by confidence and value.
The demonstration dashboard mirrors the meter schematic above, built from simulated fleet telemetry: a teal integrity LED per meter, a status stream, and a subtle amber-to-red escalation only where a tamper signature has cleared the policy-aware rules engine — never a raw accusation, always a documented, reviewable case. This is a design demonstration, not a live municipal meter fleet.
OT Lifestyle has scoped seven municipal pilot deployments for MeterTruth. Each is ready to move from architecture to a live meter fleet once onboarding support is secured with the AI platforms the design targets — OpenAI, Anthropic's Claude, and Google Cloud Platform. Securing that access is the single largest blocker between where the project stands today and a fleet-wide launch.
MeterTruth is designed as a multi-model architecture in which different AI services support detection, reasoning and infrastructure workloads at different stages of the pipeline. These are the platforms the design targets — not confirmed partnerships, sponsorships or endorsements. OT Lifestyle is seeking API access, credits and onboarding support from each provider named below.
TECHNOLOGY ECOSYSTEM · PLATFORMS UNDER CONSIDERATION — NOT CONFIRMED PARTNERSHIPS
Claude is the reasoning layer that turns a raw tamper signal into a fair, policy-checked case. This is the single most legally and ethically sensitive step in the entire platform: a wrong call here either lets real tampering slide, or wrongly accuses an indigent household. MeterTruth is built around Claude's ability to hold municipal by-law text, indigent-registration rules and a specific household's signal history in view simultaneously, and reason through them the way a careful case officer would.
Reserved for the hardest cases — where load anomalies, cover-tamper events and billing history conflict with each other. Opus reasons through the full evidence set before a case is allowed to escalate, giving revenue-protection teams a synthesis they can defend on appeal.
Runs the daily workload: dashboard summaries, case notes, resident correspondence drafts, and the fairness check that runs indigent and by-law logic against every flagged event before it can become a case.
Screens the full raw event stream from the fleet cheaply and at speed, so only signals with real substance are ever passed up to Sonnet or Opus — essential for keeping AI costs sane at fleet scale.
OpenAI's GPT models handle the parts of the pipeline that are about generating and explaining evidence, not adjudicating it — the visual and narrative layer that makes a confirmed case legible to a human reviewer, a resident, or a court. Where Claude decides whether a case is fair, GPT models are what turn that decision into a clear, reviewable document.
Converts structured tamper-event data into a plain-language incident narrative for the evidence pack, written so a non-technical reviewer, ward councillor or resident can understand exactly what was detected and why.
Generates annotated visual evidence tiles — highlighting the specific signal (magnet interference, cover event, load anomaly) on a diagram of the meter — at the volume needed to keep pace with fleet-scale case generation.
Drafts resident-facing notices and appeal guidance in English and isiZulu, so a flagged household understands its case and its right to query it, not just receives a bill.
GCP is the infrastructure the entire platform runs on — not a feature, the ground it stands on. Every signed meter reading, every model call, every case record depends on GCP services being available, fast and correctly configured, twenty-four hours a day, across seven simultaneous pilot deployments.
Trains and hosts the load-signature and tamper-pattern anomaly models that watch the meter fleet continuously, feeding confirmed signals into the Claude and GPT reasoning layers.
The fleet-wide data warehouse — consumption history, case outcomes and model training data at the scale seven simultaneous pilots require, queryable fast enough to keep the dashboard current.
Multi-region hosting and failover for the dashboard, plus Google Maps for plotting confirmed cases against the physical ward network for field-team routing and audit review.
Indicative estimates only · preliminary · subject to final scope, meter count and site conditions · not a quotation
| Phase | Scope | Indicative estimate (ZAR) |
|---|---|---|
| 10,000 meters | Retrofit or replacement, secure element, calibration | R18m – R28m |
| Platform | Dashboard, model hosting, integrations | R2.5m – R4.5m p.a. |
Every flagged event is checked against indigent registration status before it can become a billable case, so subsidised households are never mistakenly penalised.
Cases carry a timestamped, signed evidence pack — never a technician's word alone — giving residents a fair basis to query or appeal a finding.
Consumption and tamper-event data is processed under a data-sharing agreement with the municipality as responsible party, with clear retention limits.
Active-active or active-passive across AWS and GCP regions with CDN edge caching, so the dashboard stays reachable during exactly the moment it's needed most.
Circuit breakers isolate lagging or offline meters rather than blanking the dashboard; the last-known state is always shown with a clear staleness indicator.
All calls to Claude, Azure OpenAI and Vertex AI carry retry, timeout and fallback logic, with automated health checks documented against AWS and GCP Well-Architected review criteria.
Ongeziwe Tshali leads OT Lifestyle (Pty) Ltd and is the driving force behind MeterTruth, guiding the platform from concept through to a seven-site pilot rollout. The vision is straightforward: give municipalities a way to protect revenue and give households a way to prove their innocence, in the same system.
"Truth at the meter. Trust in the lifestyle."
We work directly with revenue-protection and billing teams to scope a pilot rollout, agree indigent-policy rules, and calibrate the integrity model against your own tamper history before any fleet-wide commitment.
