How to ask an assistant to analyze historic variations in my electric bills

Electric bills are a genuinely good use case for AI-assisted analysis because the data is structured and repetitive (month, usage in kWh, dollar amount) — exactly the kind of thing a small script handles perfectly and a human finds tedious to tabulate by hand. The trick is making sure the assistant actually runs code against your numbers instead of describing a trend from memory of what you typed.

Getting an accurate historic analysis

1. Export or photograph your bill history. Most utility provider portals let you download a CSV or PDF of past bills — that's a much more reliable starting point than manually retyping numbers into a chat, which introduces transcription errors.

2. Upload the file to a tool with real code execution. In ChatGPT this is the file-upload plus Advanced Data Analysis feature; in Claude it's the code execution tool. Both let the model write actual Python against your data rather than just reading it.

3. Ask for the specific comparison you want, and request a chart. "Plot my monthly kWh usage for the last 24 months and highlight months more than 20% above the yearly average" gives the model a concrete, checkable task rather than an open-ended one.

# This is the kind of code ChatGPT/Claude actually run
# against your uploaded bill CSV -- check it before trusting the output
import pandas as pd

df = pd.read_csv("electric_bills.csv")  # columns: month, kwh_used, amount_due
avg_amount = df["amount_due"].mean()
spikes = df[df["amount_due"] > avg_amount * 1.2]

print(f"Average monthly bill: ${avg_amount:.2f}")
print("Months 20%+ above average:")
print(spikes[["month", "amount_due"]])

Keep the scope honest: the code-execution sandbox has no internet connection, so it can't check whether your utility raised rates or there was a heat wave that month — it can only work with the numbers you gave it. For the "why," you'll still want to check your provider's rate-change notices or your own memory of that period alongside the AI-generated trend chart.