Home / Insights / Method
Method

A short defence of the boring audit.

The unglamorous walk-through, meter by meter, that consistently outperforms the AI-powered platform in our project records.

Ing. Papa Kusi Nimoh-Brema, SPE-GHIE · PMP ·9 December 2025 ·5 min read

It is harder than it should be to recommend a one-week walk-through with a clipboard and a serviceable power quality logger to a client who has just sat through three pitches for cloud-based, machine-learning-augmented, real-time energy intelligence platforms. The audit is the older, slower, less visually compelling option. It is also, on our project records, the one that consistently delivers more recovered cost per cedi of consultant spend than any alternative.

Three reasons — none of them a criticism of the platforms, which are genuinely useful applied to plants that have already had the boring audit done.

One — the platform does not know what is broken.

A real-time platform monitors what its sensors are wired to. It does not know about the condensate drain stuck open, because there is no sensor on the drain. It does not know the boiler air-fuel ratio has drifted, because that data lives on the boiler controller and was never federated. It does not know the chilled-water bypass has been wired open by an operator tired of resetting it, because there is no sensor within ten metres of the valve. An engineer walking the plant finds all three in an afternoon. A platform makes a well-instrumented plant smarter; it does not, by itself, make an uninspected plant inspected.

TWO INSTRUMENTS, ONE PLANT — WHAT EACH CAN SEE REAL-TIME PLATFORM sees only what sensors are wired to Main incomer kWh Metered feeders Stuck condensate drain THE WALK-THROUGH finds the physical, situational, unsensored Steam trap failed open Bypass valve wired open 8.2 bar for a 5.5 bar need The platform makes an inspected plant smarter. It cannot make an uninspected plant inspected.
Fig. 1 — Coverage comparisonBoth instruments are valuable. Only one of them finds the failed trap, the wired-open bypass, and the over-pressured network.

Two — the highest-yield findings are nearly always physical.

On the audits we have closed, recovered cost is dominated, year after year, by findings of roughly this kind:

None require artificial intelligence. They require somebody to look. Cumulative recovery from a serious walk-through on a mid-sized facility is, in our experience, in the [PERCENTAGE] band of annual energy spend, payback inside [TIMEFRAME] on the fixes and roughly [MULTIPLE] return on the audit fee — a one-off recovery, after which the platform (or disciplined sub-metering) earns its place keeping the gains.

Three — the audit teaches the plant something the platform does not.

The most underrated output is not the report. It is the two days the auditor spends walking with the plant's own engineers, asking why the bypass is open, what the receiver pressure has been, when the boiler was last combustion-tested. They know most answers, or can find them by lunchtime. What they have rarely been asked is to think about their plant as an energy system rather than a sequence of fires. The boring audit transfers a way of looking; the plant keeps it after the consultant leaves. The platform transfers a dashboard.

So when is the platform the right answer?

After the boring audit, not instead of it:

The principle. In that sequence every layer earns its capital, and the order of the layers is everything. The boring audit is not an outdated method. It is a foundational one. Skip it, and the more impressive engineering above it has nothing to stand on.

A note on figures. Specific quantities in square brackets are placeholders for values drawn from PKNB Consult engagement records. They are completed against project documentation prior to client circulation; client identities are withheld throughout as a matter of professional confidentiality.
Ing. Papa Kusi Nimoh-Brema · Principal Consultant, PKNB Consult · Accra, Ghana