Everything in large text is what you say out loud. Everything in an orange box is what you do with the laptop. You do not need to memorise it — read it, say it in your own words, and keep to the order.
The single most important rule: do not oversell. The strongest thing about this proposal is that it admits what it cannot do. That is what will make them believe the rest.
index.html, then click through to the console once, before anyone arrives.Good morning. My name is Prince Owusu. I run POA Nexgen Technologies, a Ghanaian software company here in Accra.
I am going to show you something we built, and I am going to be very straight with you about what it does and what it does not do. It will take about fifteen minutes.
Here is the problem in one sentence. Right now, when a transformer fails somewhere in Ghana, the first person who knows is the customer. Not the engineer. Not the utility. And not you.
Pause. Let that sit.A transformer is the metal box on the pole that takes the big electricity coming down the road and turns it into the electricity that goes into people's houses. There are tens of thousands of them across the country.
Almost none of them can tell anyone how they are doing. They sit there quietly until something goes wrong. Then somebody's fridge stops, and they call. That call is our early warning system today.
Now here is the part that costs the most money, and it is a small thing that most people never think about.
When power goes off on one line, there are two completely different reasons it can happen. One is a fuse at the bottom of the pole. A man can stand on the ground and change it in ten minutes. The other is a fuse at the top of the pole, on the high voltage side. For that, a crew must climb, with different tools and different safety gear.
From the ground, both look exactly the same.
Pause.So the crew drives out with the wrong equipment. They cannot fix it. They drive back. They come again tomorrow with the right thing. Meanwhile the fault is still there, and while it is still there it is quietly burning out people's motors, their fridges, their machines — because the power is not off, it is half on, which is worse.
That is one fault type. There are others: transformers carrying more load than they were built for, fuses being stolen, boxes being opened at night for the copper and the oil.
And underneath all of that is your problem, not theirs. When you ask a utility how long the power was off, the only answer you have is the one they give you. There is no independent record. You are regulating using numbers reported by the people you are regulating.
So we built GridWatch. It is two things.
First, a small weatherproof box that clips onto a transformer that is already there. It does not replace anything. Nobody's power goes off while we install it — the sensor is a clip that closes around the outside of the cable, like a clothes peg. Nothing is cut.
Second, the software. It watches every box, works out what is actually wrong, and tells the right person, with the right instruction, before the phone starts ringing.
Let me just show you.
This is a simulation — there is no box on a pole yet. But the thinking behind it is real, and it is running live right now.
Every dot is a transformer. Green is healthy, amber is a warning, red is a serious fault. Right now this is one district, but you can change it here —
— and the same system covers every district, every region, and every operator: ECG, NEDCo, GRIDCo. If a utility engineer signs in, they see only their own transformers. If you sign in, you see all of them. That is the difference between being told and being able to check.
Now the important part. Slow down here.Let me break one. This is the fuse at the bottom of the pole.
Look at this line. It says holding, one of three. Now two of three.
The system has seen the fault. It is not telling anybody yet.
This matters more than anything else I will show you today. These sensors are not perfect — no sensor is. If we shouted every single time a reading looked odd, we would be wrong often, crews would drive out for nothing, and within one month every engineer in the country would stop believing this system. So it waits. It has to see the same fault three times in a row before it says a word.
Now it is sure, and now it speaks. And look at what it says. Not "there is a problem." It says: open phase, low voltage fuseway. Ground level. Carry a J-type fuse link.
Now watch what happens when I make the other fault — the one at the top of the pole.
Same transformer. Same street. Completely different answer: single phasing, high voltage cutout. Climb to the primary. Do not open the low voltage fuseway.
That one sentence is the difference between a job done this morning and a crew coming back on Thursday.
And this is what the engineer in the field actually holds. The job, the readings at the pole, and the list of what to bring. It works when there is no network — it saves and sends when the signal comes back. If there is no data at all, it goes by SMS.
One last thing on this screen. When we fit a box, the engineer checks it against a proper meter and enters the reading. That single step is what makes every number in this system trustworthy. Without it, we would be guessing accurately-looking numbers. With it, we know how wrong we can be — about two percent.
And this is yours. Outage minutes, measured by the equipment, not reported by the utility. This is the number you regulate with.
Now I want to tell you three things that most people selling you software would leave out. I would rather you hear them from me today than find them yourself in six months.
Slow down. This section is why they will trust you.Number one. This system cannot catch electricity theft. Not yet.
To catch theft, you compare what the transformer sends out with what all the customers actually used. That needs a working meter on every house that can be read remotely. That does not exist yet in most places. So we will not sell you something we cannot deliver. When the meters are there, this system is ready for it — but today, no.
Number two. On its own, this does not pay for itself.
On our own numbers, what we save by preventing transformer failures is less than what it costs to run. I am not going to stand here and promise you a payback in seven months, because it is not true. What you get for the money is the ability to see your own network, and safety, and an independent record. Those are worth paying for. But I will not dress them up as profit.
Number three. I have to declare something.
Earlier in this process, my company drafted terms of reference for this kind of system. That means if you now put this out to tender using that document, we would have written our own exam. So we are asking you not to use it. Put out your own specification, tender it openly, and we will compete like everybody else.
Pause. Let them react.I am telling you this now because a contract that cannot survive an audit is worth nothing to me and less than nothing to you.
So here is what I am not asking for. I am not asking you to buy a national system. Nobody should, and I will tell you why.
Nobody in this country currently knows how often distribution transformers actually fail. Not you, not ECG, not us. Everyone has an opinion. Nobody has a measurement.
So what I am asking for is this. One hundred and fifty transformers in one district, fitted with these boxes. And another one hundred and fifty, chosen to be as similar as possible, with nothing fitted at all.
Eighteen months later we compare the two groups. Then, for the first time, you will have a real number instead of everybody's opinion. And whatever that number says — even if it says this was not worth it — you will own that answer and be able to defend it to the Auditor-General.
The cost of that is six hundred and seventy-two thousand US dollars over eighteen months. That is everything: the boxes, the engineering, the software, the installation, the training, and an independent evaluation that we do not control.
And there is one more thing I would ask, which costs nothing.
In our proposal there are fourteen numbers marked with a small dagger. Every one of them is a guess we have labelled as a guess. Give us one meeting with your technical people and ECG, and we can turn most of those guesses into your own real figures. Then the proposal in front of you would be built on your numbers instead of mine.
And there is one more thing I would ask for, which costs you nothing at all.
You are already replacing transformers across the country. Every single one of those is a chance to fit monitoring at a fraction of the price.
Think about why. When a transformer is being replaced, the crew is already there. The vehicle is already there. The power is already off, because the outage is happening anyway. And the new transformer is sitting on the ground, switched off, before it goes up the pole.
Fitting our box at that moment costs about a third of coming back to do it later. No separate trip. No working on live wires at all. And not one extra minute without power for anybody.
Pause here. This is the offer they can accept today.So I am not only asking you to fund a pilot. I am asking for a decision that costs nothing: that from now on, every new transformer you install gets fitted on the way in. Retrofit is for the fleet you already have. The ones you are buying anyway should arrive ready.
Let me finish where I started.
Today, when a transformer fails, the customer knows first, the utility knows second, and you find out last — from a report written by the people you regulate.
What we are offering is simply this: that you find out first, and from your own equipment.
Everything I have shown you is in the proposal, including the three uncomfortable parts. I have brought copies. Thank you for your time — I am happy to take any questions.
Short, honest answers. If you do not know, say you do not know and offer to find out — that is far stronger than guessing in this room.
"Is this actually built, or is it just a demonstration?"
The thinking is built and tested. The box on the pole is not. The software that decides what has gone wrong is real and working — you just watched it. The hardware needs about five to seven months of engineering before the first unit is on a pole. I would rather say that clearly than let you find out later.
"How do you install it without switching off the power?"
The sensor is a clip that closes around the outside of the cable. It reads the electricity through the cable's magnetic field, without touching the metal inside. Nothing is cut and nothing is disconnected. The crew is certified for live-line work and wears rated protective equipment.
"How do we know your readings are correct?"
Two ways. First, when we install each box the engineer checks it against a proper reference meter and stores the correction — you saw that screen. Second, we re-check a sample of the fleet every year. That is why we say we are accurate to about two percent, rather than just saying "accurate".
"Can we use this data to fine a utility?"
Not on its own, and we would advise against it. These are good industrial sensors, not certified revenue meters. What this data does is tell you where to send an inspection. The inspection produces the evidence. If you fine somebody using only this, the first competent lawyer will overturn it and the whole system loses credibility with it.
"ECG will say this is their equipment and their network. How does that work?"
We recommend the boxes are owned and operated by the utility, under a technical directive from the Commission, with a guaranteed read-only feed to you. That way the equipment sits with the people who already carry the safety responsibility for the pole — and ECG has no reason to resist it, because it helps them too.
"Why is this so much more expensive than the earlier figure we saw?"
Because the earlier figure was wrong, and I would rather correct it now than discover it during delivery. It left out the hardware engineering completely, priced the installation at a third of what a certified crew actually costs, and assumed a tax exemption we probably do not qualify for. The corrected number is bigger, but it is one we can actually deliver at.
"What happens to the boxes if the pilot fails?"
They stay yours — you have paid for them and they keep working on those 150 transformers. And the evaluation still has value: knowing that monitoring does not pay on Ghana's network would itself be worth having, because it would stop the same money being spent again at ten times the scale.
"Is our data safe, and where does it live?"
The boxes talk to us over a private mobile channel that cannot reach the public internet, and everything is encrypted. Where the data physically sits is your decision, not ours — our own cloud, the government's NITA data centre, or your own servers. We would like a written instruction from you on that early, because it changes how we build.
"Who else is using this?"
Nobody yet. This would be the first deployment, which is exactly why we are proposing a controlled pilot rather than a national contract. I would be worried if somebody told you a first-of-its-kind system should go straight to 5,000 units.
"Are these on poles or on the ground? Does that change anything?"
Both, and yes it does. The smaller ones on poles need elevated work, so they cost more to fit. The bigger ones on concrete inside cages are cheaper and safer to reach, but the metal blocks the signal so they need an external aerial. Our survey establishes the mix in your chosen district before we order anything — the installation figure in the quotation is a blended rate, not a per-pole price.
"How many transformers is this replacement programme actually installing?"
I do not know, and I would not want to quote you a number I read in the newspaper. That is one of the figures we have asked to confirm with the Ministry and ECG directly. But whatever the number is, fitting those on the way in is cheaper than coming back for them later — and that holds whether it is five hundred or five thousand.
"Can you do it cheaper?"
We can do it smaller — fewer units, shorter pilot. We should not do it cheaper by cutting the surge protection, the calibration or the evaluation, because those are the three things that decide whether the answer at the end is trustworthy. I would rather reduce the scope than reduce the quality of the answer.