Bill too high? How to use AI to find out where you consume the most and save money

The latest electricity bill is 120 euros, while the previous one was 90 euros. If we ask an AI-based chatbot (e.g. Gemini, ChatGPT etc.) “How can I save?”we will probably get advice that we already know. Instead, let’s try asking him to compare the two bills, so as to understand if we have consumed more energy, if the price of electricity has changed or if an adjustment appears in the last invoice. Based on the response we get we can understand where it is best to intervene. In fact, if the increase concerns the price, it is worth looking at the contract, otherwise if the kilowatt hours (kWh) consumed have increased, we can go and look for the days in which it happened. And finally, if we are thinking of changing a habit, for example running the dishwasher in the evening, we can calculate how much it would save us before even trying.

The consumption, tariffs, temperatures and times of the examples are hypothetical: obviously each user can think of re-proposing these questions and the calculations with their own home data.

Do you consume more on the weekend? Consumption comparisons

Suppose that two bills cover periods of similar duration and that the actual readings confirm an increase in consumption: the total tells us how much more we have consumed. To understand when, we can retrieve daily consumption from the supplier or distributor portal, if available. In this case, we could think of providing the AI ​​chatbot with a table with two columns, date and kWh consumed, and a precise request:

Compare consumption over the last eight weeks. Compare weekdays to weekdays and weekends to weekends. Since when did consumption increase? Tell me the dates and values ​​on which you base your answer.

Let’s imagine that from Monday to Friday the house remained around 8 kWh per day, while in the last four weekends, the average consumption on Saturday and Sunday rose from 9 to 14 kWh per day. Those extra 5 kWh, repeated for eight days, explain the approximately 40 kWh of additional consumption. At this point the question becomes more limited: what have we started doing on the weekend that we didn’t do a month ago?

Maybe we remember that, since it’s colder, on Saturdays and Sundays we turn on an electric heater in the study. For example, a 2 kW heater that works at full power for two and a half hours consumes exactly 5 kWh: from here we would realize that the hypothesis adds up to the numbers. If we also have hourly data, we can check whether the increase coincides with the hours in which we use the room. The meter measures the consumption of the whole house: the chatbot can identify the days and times to examine, while we connect them to what happens in the study.

The opposite case could also emerge, i.e. similar daily consumption and higher spending. In that case we can ask the AI ​​assistant to combine the amounts of the two invoices, item by item. This distinction is important because the bill includes a portion linked to consumption, a fixed portion and one linked to the power used, therefore an increase in the total can have different origins.

Is it necessary to turn on the heating at six in the morning?

We want it to be 20°C in the living room at eight in the morning, so we’ve programmed the thermostat to start heating at six. Some mornings, however, it happens that the room is already warm at seven. Can we start the heating a little later without having breakfast in the cold? In this case, you might consider writing down the outside temperature for a few mornings, the temperature of the living room when you turn it on, the start time and the time the room reaches 20°C. If the thermostat keeps a history, so much the better: we can use the recorded data directly. Then we ask the chatbot:

I want to find 20°C at eight in the morning. Compare mornings with similar starting and outside temperatures. Which days was the living room hot well in advance? Could you suggest a small switch time change for me to try and then tell me how to check it?

On three mild mornings, we imagine the living room reached 20°C between 7:05 and 7:20 am, while on the coldest morning it reached it at 7:55 am. A first attempt might be to delay the start by 15 minutes on a mild day and check the temperature at eight. However, the 7:55 data is equally valuable: in the cold morning we had just five minutes of margin.

The result also depends on how the house is heated and the type of terminals installed. In fact, a radiant floor takes time to heat up and continues to release heat after it is turned off, while a split introduces hot air into the room after just a few minutes. The temperatures recorded in the house help to choose a time suitable for our system, the economic benefit then remains to be verified. The thermometer tells us if we have maintained comfort at eight o’clock and to estimate the savings we must also compare consumption, observing several days with similar weather and habits.

When is it best to start the dishwasher: evening costs

It often happens that you are advised to start the dishwasher at night, but how much is this change of habit worth in euros with the tariff stipulated in your electricity supply contracts? To find out, all you need is the program consumption, the number of washes and the price we pay in the different time slots. We could therefore think about asking the AI ​​chatbot:

My dishwasher consumes 1 kWh per cycle and about 20 washes are started per month. I provide you with the prices of the ranges of my contract. How much would I save in a month if I changed my washing schedule? Show me the calculation.

Suppose that 1 kWh costs 30 cents in the range in which we usually start the dishwasher and 25 cents in the evening. By moving all the washes we would save 5 cents per cycle, that is, one euro per month. If the price per kWh is the same at all hours, the tariff advantage is zero. Knowing how much we can save helps us understand if it is worth changing time and if the new time slot fits in with our habits.

If we have a photovoltaic system, the situation is different: it might in fact be convenient to start the dishwasher during sunny hours. Suppose that at midday the energy produced by the surplus panels is enough for a washing cycle, which requires 1 kWh. If putting that kWh into the grid earns us 4 cents and buying it in the evening costs us 25, by starting the dishwasher at midday we avoid spending 25 cents and give up 4 cents in revenue. The advantage of self-consumption is therefore 21 cents per wash. To redo the calculation, simply replace the prices in the example with the cost per kWh of the electricity we purchase and the compensation for each kWh we feed into the grid.

The chatbot is also useful when the calculation shows a saving of just one euro per month. Maybe we decide to continue running the dishwasher at the time that is most convenient for us and to concentrate on the consumption of the study on Saturday afternoon. If the next bill is higher than expected, we can ask him:

On which days did I consume more than usual and what changed in the house just then?

don’t miss this article
Is electricity cheaper during the day than at night? Not for everyone: how it works on the bill

Sources

ARERA, guide to the fixed and variable components of the GSE electricity bill, criteria for valorising the electricity fed into the grid.