Check whether a half-hourly electricity tariff would be cheaper by giving Claude 87 days of smart meter readings

Posted

Apr 7, 2026 by Beckie Brown (Notes & Sparks)

·

Last checked

Oct 19, 2026

·

Method:

How we verify

Evidence: Screenshots or logsThe source shows screenshots, a transcript or the agent's output.

Outcome

Beckie Brown's household was on a fixed Octopus Energy tariff at 26.90p per kWh and wondered about switching to Agile, whose price changes every half hour. They downloaded 87 days of half-hourly smart meter readings, Octopus's published historical Agile prices for their region and their own tariff details, and asked Claude to work it out from the real data. In about 30 minutes Claude matched usage to prices, projected a full year from three years of Agile history, and produced an 8-page report: about £171 a year saved with no changes, around £218 with small habit shifts, a peak penalty of about £133 from 4 to 7 pm, and no exit fee. They switched. In a follow-up post on 12 May 2026, the first 30-day bill was just under £78, against about £120 for the same usage on the old tariff, in a record month for negative prices.

Agent

Claude

The source didn't share its exact prompt, so we wrote this from what it describes.

Run this task now, exactly as written below. Everything you need is here, so start straight away and only stop if something required is missing or a rule below tells you to.

Goal: Work out whether a time-of-use electricity tariff would save my household money, using our real usage.

Fill in before running:
- My smart meter readings export (half-hourly or hourly): [UPLOAD]
- My current tariff (unit rate, standing charge, end date, exit fees): [DETAILS]
- The tariff I'm considering and where its past prices are published: [DETAILS]

Remove account numbers and your address from any files before uploading.

Rules:
- Use my actual readings matched to the actual past prices; say where you had to estimate.
- Show the result by hour of day and by season, not just one total.
- Spell out the risks: peak-hour costs, price spikes and caps, and exit fees.
- Show your working so I can check it.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.

Output: A short report with the yearly difference, an hour-by-hour chart, risks and a clear recommendation.

You'll know it worked when: I can decide in minutes and know when to shift usage.

Run this task now, exactly as written below. Everything you need is here, so start straight away and only stop if something required is missing or a rule below tells you to.

Goal: Work out whether a time-of-use electricity tariff would save my household money, using our real usage.

Fill in before running:
- My smart meter readings export (half-hourly or hourly): [UPLOAD]
- My current tariff (unit rate, standing charge, end date, exit fees): [DETAILS]
- The tariff I'm considering and where its past prices are published: [DETAILS]

Remove account numbers and your address from any files before uploading.

Rules:
- Use my actual readings matched to the actual past prices; say where you had to estimate.
- Show the result by hour of day and by season, not just one total.
- Spell out the risks: peak-hour costs, price spikes and caps, and exit fees.
- Show your working so I can check it.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.

Output: A short report with the yearly difference, an hour-by-hour chart, risks and a clear recommendation.

You'll know it worked when: I can decide in minutes and know when to shift usage.

Run this task now, exactly as written below. Everything you need is here, so start straight away and only stop if something required is missing or a rule below tells you to.

Goal: Work out whether a time-of-use electricity tariff would save my household money, using our real usage.

Fill in before running:
- My smart meter readings export (half-hourly or hourly): [UPLOAD]
- My current tariff (unit rate, standing charge, end date, exit fees): [DETAILS]
- The tariff I'm considering and where its past prices are published: [DETAILS]

Remove account numbers and your address from any files before uploading.

Rules:
- Use my actual readings matched to the actual past prices; say where you had to estimate.
- Show the result by hour of day and by season, not just one total.
- Spell out the risks: peak-hour costs, price spikes and caps, and exit fees.
- Show your working so I can check it.
- Check facts that can change (prices, opening hours, dates, availability, rules) against the official source, and tell me anything you couldn't verify.
- After a month, compare my real bill with what the old tariff would have cost.

Output: A short report with the yearly difference, an hour-by-hour chart, risks and a clear recommendation.

You'll know it worked when: I can decide in minutes and know when to shift usage.

Before you finish: re-read my goal and constraints, check your answer against each one, and list anything you assumed or couldn't confirm.

Source

https://notesandsparks.substack.com/p/how-ai-helped-me-switch-energy-plans