International manufacturer of modular cold rooms and cleanroom envelopes, commercially accountable for the thermal performance of a large installed base it does not own, occupy or manage.
640 Cold Rooms. 180 Sites. Fourteen Months to an Enterprise Service Moat.
How do you defend a maintenance contract on equipment you built, but no longer see?
The manufacturer absorbs the cost of third-party operator door habits, unrecorded power dips and deferred customer maintenance, paying for each through eroded warranty margins. Each year the service agreements go back out to tender against regional mechanics competing on hourly labour cost alone, and renewals hover at 71%.
Pragmatos 3.0 goes in on a 40-room reference group first: read-only edge gateways, per-room energy attribution, and drift baselines that separate mechanical wear from loading habits. Confirmed drift becomes a human-approved Orbis service playbook, and every recommendation and approval lands in a timestamped operating record the manufacturer puts on the table at renewal.
The customer class is an international engineering manufacturer specialising in modular cold rooms, blast freezers, insulated panel assemblies and cleanroom envelopes, delivered under comprehensive design-build-maintain agreements. It serves dairy, seafood, perishables and controlled storage, and its core engineering discipline is thermal preservation.
Internal manufacturing operations across its own factories are optimised. The existential financial risk resides almost entirely beyond them: continuous commercial liability for thermal performance across deployed installations it does not own, occupy or manage.
Without continuous data from those rooms, the manufacturer absorbs the costs of third-party operator door habits, unrecorded power dips and deferred customer maintenance, paying for each through eroded warranty margins.
Field engineers spend hundreds of hours in transit across borders only to diagnose a worn door seal, a frozen evaporator coil or a clogged condenser that continuous operational telemetry could have surfaced days earlier.
Equipment is sold once, after which annual service agreements are re-tendered against regional mechanics competing solely on hourly labour cost. Without a continuous operational record of what the equipment delivered, renewals hover at 71%.
Four structured phases over 14 months, proving the unit economics on the 40-room reference group before the rollout scales across the portfolio.
Every room in the pilot group is surveyed for what its controller actually exposes over Ethernet. Rooms whose read path cannot be confirmed on site are excluded upfront, which sets a reliable hardware boundary before anything is installed.
Read-only edge gateways go in with roughly 72 hours of local buffering. Compressor, evaporator, door and thermal variables are mapped into one semantic namespace, and per-asset energy attribution is live by week 11.
Pulse calibrates an individual thermodynamic response curve for each room, to separate loading habits from mechanical decay. Orbis service playbooks are written with a designated, named approver for every ticket.
The rollout scales toward 640 cold rooms across 180 customer facilities in 11 countries. Telemetry becomes a priced, tiered service contract that customers renew to keep a continuous operating record of every room.
Figure 2 · Human-Approved Service Chain: Pragmatos 3.0 has no write path into plant equipment, so nothing in this chain acts on the room by itself. Telemetry triggers a Pulse drift evaluation, which becomes a structured Orbis recommendation. A named engineer approves or rejects it. Both the recommendation and the decision are logged, creating an auditable record that makes the maintenance contract defensible at renewal.
There is no write path into the refrigeration equipment, and no remote override switch exists anywhere in the platform. That is what shortens a food-safety or plant IT review: there is nothing left to argue about. The local edge buffers roughly 72 hours of telemetry, so a network outage costs continuity, not data.
The aggregate utility meter is broken down into room-by-room consumption by cross-referencing compressor draw with suction pressure, ambient temperature and door cycles. An inefficiency such as a degraded seal driving a runtime surge becomes an actionable line item instead of hidden operational overhead.
Threshold alarms fire falsely whenever loading dock doors stay open during morning logistics. Pulse separates genuine mechanical wear from operator activity by correlating thermal response curves. Confirmed drift is routed as an auditable recommendation with a named engineering sign-off.
A manufacturer in this position already has threshold alerts, and has learned to ignore most of them. The value worth paying for is not another alert dashboard. It is placing a continuous, auditable operating record on the table at contract renewal, which takes the renewal off labour price and presents a performance history that a third-party maintenance provider does not hold.
A temperature threshold asks whether the room is already failing. A per-room energy baseline asks whether the machine behind it has started to work harder for the same result. The second question is answered days before the first.
Figure 1 · Drift Detection vs. Lagging Thresholds: Under a legacy alarm, chamber temperature stays within safe bounds until the compressor fails at Day 0, which means an emergency call and perishable stock at risk. Pulse watches power draw relative to delta-T, and the energy curve leaves its baseline eight days earlier. That is enough time for a scheduled, low-cost component swap in a regular maintenance window. Pulse raises the question; a named engineer decides.
| Before | With Pragmatos 3.0 (modelled) |
|---|---|
| One blended utility bill; no thermal breakdown per room. | Per-asset runtime and energy, hour by hour. |
| First warning of a fault is the customer's phone call. | Drift surfaced days ahead of the thermal excursion. |
| Engineer travels to diagnose, then returns to repair. | One visit, correct part, 88% first-time fix. |
| Threshold alarms ignored after too many false calls. | Mechanical drift separated from operator loading behaviour. |
| Warranty settled by argument over several days. | Settled from the timestamped record in minutes. |
| Service contracts re-bid on price. 71% renewed. | Renewed on verifiable evidence. 94% renewal target. |
Each result on page 1 is either an Assumption (a rate we chose to model, stated so you can change it) or Derived (arithmetic on those assumptions). Change an assumption and the derived figure moves with it.
| Figure | Type | Inputs and working |
|---|---|---|
| €540,000 | Derived | Refrigeration energy baseline: 40 rooms × 12 kW × 14 h/day × 365 d × €0.22/kWh = €539,616 ≈ €540,000/yr |
| 6.2% | Assumption | Share of baseline recovered once per-room attribution exposes gasket, condenser and defrost faults |
| €33,480 | Derived | 6.2% × €540,000 = €33,480/yr |
| −34% | Assumption | Share of dispatches avoided when faults are pre-triaged remotely |
| €18,360 | Derived | 40 rooms × 3 calls/yr = 120 dispatches × €450 = €54,000; 34% (≈41 calls) = €18,360/yr |
| €51,840 | Derived | €33,480 + €18,360 on the 40-room group, before any warranty saving |
| 0 | Assumption | Design target: wear caught as a planned swap (≈€2,400) instead of an emergency overhaul plus stock loss, roughly ten times that on one loaded room |
| Days→min | Assumption | The record exists before the dispute does, so adjudication is a lookup rather than an investigation |
| 61%→88% | Assumption | First-time fix when the engineer arrives with the fault signature and the part |
| 71%→94% | Assumption | Renewal rate the blueprint is built to reach; 71% is the typical re-bid starting point |
| Read-only edge, ~72 h buffer, no write path |
| Pragmatos Live: ingest, namespace, dashboards, rules |
| Pulse: per-room energy and drift baselines |
| Orbis: rules a named engineer approves |
| Timestamped record, renewal reporting |
Industrial IoT deployed as passive telemetry is treated as overhead and cut during budget reviews. Pragmatos 3.0 turns operational telemetry into an institutional asset: an auditable record of recommendations, named approvers and recorded outcomes. That record is what the manufacturer presents at contract renewal, what resolves a warranty dispute in minutes, and what a low-cost local competitor cannot produce.