Loading…
AI for data analysis: from spreadsheet to insight
6 minutes · 8 steps · workout 4 of 6 in this module
What you'll be able to do
By the end of this workout, you'll be able to hand a spreadsheet to an AI assistant with the right context, catch the mistakes it makes most often, and turn its summary into a decision you've verified yourself.
A first look: What an AI assistant can and can't do with your spreadsheet
Paste or upload a campaign export and an assistant like ChatGPT, Gemini or Claude can clean it, summarise it, write the formulas, spot patterns and draft the chart in minutes. What it can't do: know what your columns mean unless you say so, know which metric matters for this decision, or check whether the data itself is right. It also tends to answer confidently either way. So the job splits: the assistant does the arithmetic and the first read, you set the question, supply the meaning and verify the numbers that will drive a decision.
Example: Northpack's weekly ad export has a column called “results.” In the Meta rows it means purchases; in the TikTok rows it means link clicks. An assistant that isn't told this will compare them as if they were the same thing.
What this workout covers
- What an AI assistant can and can't do with your spreadsheet
- Case: “TikTok is your best channel”
- The analysis prompt: dictionary, question, metric, formulas
A question from this workout
You paste a GA4 export into an AI assistant and ask “What's interesting here?” It replies: “Traffic grew 30% month over month, driven by organic search.” What do you do first?
- Paste the sentence into the weekly report, it sounds right
- Ask it which columns and rows it used and what the two monthly totals were, then check those two totals in the sheet yourself
- Ask it for five more interesting findings to fill the report
- Reject it, since AI analysis can't be trusted
Pick your answer first, then open the reasoning below.
Show the answer and the reasoning
Answer: B. Ask it which columns and rows it used and what the two monthly totals were, then check those two totals in the sheet yourself
“Show your work” is the single most useful analysis prompt. Two totals take a minute to verify and tell you whether the 30% is real. The most common mistake is the first option. A confident sentence isn't evidence, and once a number is in a report, people repeat it without the caveat. Rejecting everything (option d) throws away the speed for no gain; verifying two numbers keeps both.
The other steps work the same way: you decide first, then see why each option is right or wrong.
How you practice here
- 2 short concept cards
- 2 multiple-choice questions
- 1 campaign case with real-looking numbers
- 1 put-in-order exercise
- 1 guided calculation
- 1 recall question from an earlier workout
Part of the module: AI in Marketing
Make AI your teammate for content, ad creative, analysis and research, and understand how AI search is changing marketing.
More workouts in this module
- Writing good prompts: role, context, examples, format
- Creating content with AI: speed, brand voice and control
- AI for ad images and video creative
- AI search: how ChatGPT, Gemini and AI Overviews are changing traffic
- Risks: hallucinations, copyright, privacy and brand safety
Programs that include this workout