I Was Impressed by the Model. The Manager Wanted Proof.

I have spent a while building a model of a pumping station. It has grown into something I am proud of. It runs the pumps, it tracks the energy, and it has let me publish a couple of conference papers on some new techniques.

I delivered a demo to the people who run a plant and I showed them what it could do.

I could tell within a few minutes that it was not landing how I had hoped.

The room was polite and the modelling was sound. But the thing I was excited about was not the thing they were interested in.

I brought them a model that can tune a controller offline, before anyone touches the real loop. A model that can connect to the plant's PLC and let you commission the control system against a simulated plant before the site is even live. A model that can sit alongside the process and let an optimiser search thousands of pump schedules with no risk to production.

To me, each of those is genuinely useful. I had put real effort into making them work.

The manager did not ask about any of them.

He asked one question, in different forms, several times. Will this save me money on the energy? And can you show me that it does.

That was it. Not the method. Not the elegance of running an optimiser against a twin. Not offline commissioning, which to him solved a problem he did not have. He was standing in front of a station that costs a known and large amount to run every year, and he wanted to know whether this would take a measurable slice off that number. Ten percent a year would have earned his attention. The capabilities on their own did not.

I am an engineer, not a salesperson. I do not think the pitch failed because I delivered it badly. It failed because I was presenting the wrong thing. I offered a piece of software. They needed a solution.

I think this is a particular blind spot that engineers have. We fall for the potential of a technology. We see the range of what it could do, the cleverness of the approach, and we treat that potential as if it were obviously valuable to everyone else. A manager does not buy potential. A manager buys a benefit he can count, on the plant he is accountable for.

The feedback I got was short. Prove it.

I appreciated that far more than any polite enthusiasm would have given me. It was honest, it was fair. A model that might save energy is a curiosity. A model that has been shown to save energy on this station, measured against this station's own data, is a tool worth paying for. He was right to want the second one.

What I walked out with was not a sale. It was permission to bring the simulation model onto the live plant, develop it into a digital twin, and demonstrate the energy benefit for real.

The difficult part is not building a model that can optimise a plant. It is earning the right to connect that model to a real plant.

Before a plant operator will trust it, the model has to match their plant, use their data and demonstrate a result they can verify. In this case, that result is energy saved.

That proof is not separate from the engineering. It is the work that turns a capable simulation into something the plant can actually use. Establishing the baseline, calibrating the model, testing its recommendations and measuring the outcome are all part of building the system.

I am not going back to make the model more impressive. It is already capable of more than I managed to communicate in that room. The next step is tighter and more difficult: use the technology to develop a solution on one real pumping station and make the saving visible in that station's own numbers.

If it can produce a repeatable reduction that the plant can measure for itself, the model has earned its place. If it cannot, none of its sophistication matters.

I walked into that meeting thinking I needed to explain what the model could do. What I needed was evidence of what it was worth.