Prompts for customer data analysis
This resource is based on The AI research stack behind a solo founder's Nordstrom deal, featuring Marianna Sachse of Jackalo, published on the AI Lab by ActiveCampaign.
Get the prompt template
This prompting guide is based on The AI research stack behind a solo founder’s Nordstrom deal, featuring Marianna Sachse of Jackalo, featured on the AI Lab by ActiveCampaign.
How to use these prompts
These prompts are based on Marianna Sachse’s workflow for analyzing customer calls, survey responses, and Shopify data as a solo founder. Copy, paste, and fill in the brackets with your specifics. Sachse runs most of these across multiple AI tools (ChatGPT, Claude, Gemini) and cross-references the outputs. If two tools agree, she’s confident; if they diverge, she digs deeper.
Prompt 1: Customer call analysis
- Best for: Pulling key learnings and action items from a recorded customer conversation
- Use with: ChatGPT or Claude (Sachse uses both and compares)
Customer call analysis
Here is a transcript from a customer call. I sell [PRODUCT/SERVICE] to [CUSTOMER DESCRIPTION: e.g., parents of kids ages 4–12 who care about sustainability].
Please analyze this conversation and give me:
- The top 3–5 key learnings about my customer’s needs, frustrations, or motivations
- Any patterns you notice in how they describe their problem or their decision to buy (or not buy)
- Specific action items I should take based on what they said
- Any quotes that stand out as particularly revealing
Transcript:
[PASTE TRANSCRIPT]
Variables to fill in:
- [PRODUCT/SERVICE]: what you sell
- [CUSTOMER DESCRIPTION]: who this customer is and why they matter to you
- [PASTE TRANSCRIPT]: export from Fathom or your call recorder
What to expect: A structured summary with themes, standout quotes, and a short action list. Run this in two different tools and compare, e.g. look at the gaps between outputs often surface the most interesting insights.
Follow-up prompt
What did this customer say (or imply) about [SPECIFIC TOPIC: e.g., price sensitivity, how they found us, what almost made them not buy]? Is there anything in the transcript that speaks to [BUSINESS QUESTION]?
Prompt 2: Survey response analysis
- Best for: Synthesizing patterns across a batch of customer survey responses
- Use with: ChatGPT (handles CSV uploads well)
Survey response analysis
I’m uploading results from a customer survey. Here’s the context:
- What I was trying to learn: [YOUR RESEARCH GOAL: e.g., why customers return items, what almost stopped them from buying]
- Who I sent this to: [AUDIENCE: e.g., customers who purchased in the last 90 days]
- Business decision I’m trying to make: [DECISION: e.g., whether to expand our size range, how to improve the checkout experience]
Please analyze the responses and tell me:
- The main themes and patterns in what respondents said
- Any surprising or unexpected findings
- The most important things I should act on
- Whether any groups of respondents seem to have meaningfully different experiences or needs
[ATTACH CSV EXPORT FROM GOOGLE FORMS OR TYPEFORM]
Variables to fill in:
- [YOUR RESEARCH GOAL]: what question you were trying to answer
- [AUDIENCE]: who received the survey
- [DECISION]: the business decision this is meant to inform
What to expect: A thematic breakdown with actionable priorities. Sachse cautions that AI “sometimes focuses on what it thinks is most important and will miss a whole part of the survey.” Skim the raw CSV and ask about anything the AI didn’t address.
Follow-up prompt:
You focused mainly on [TOPIC]. I also need you to analyze the responses to [SPECIFIC QUESTION OR SECTION]. What do those responses reveal?
Prompt 3: Survey improvement
- Best for: Improving your next survey based on what the current one revealed (and didn’t)
- Use with: ChatGPT or Claude (after running Prompt 2)
Survey improvement
Based on the analysis we just did and the responses you reviewed, where could I be asking different questions? Specifically:
- What questions would get me more useful or specific answers?
- Are there topics my customers clearly care about that I’m not asking about directly?
- Are any of my current questions ambiguous or leading in a way that skews responses?
- What’s one question I should add and one I should cut?
Variables to fill in: None. Run this directly after Prompt 2 in the same conversation.
What to expect: Concrete suggestions for rewriting or adding questions. This is a tactic Sachse uses to compound the value of every survey she runs, each one improves the next.
Follow-up prompt
Write a revised version of the survey with those changes applied. Keep the same overall goal but make the questions tighter and more specific.
Prompt 4: Shopify data analysis
- Best for: Getting answers to specific business questions from Shopify reports
- Use with: Claude (Sachse finds it especially strong for data analysis)
Shopify data analysis
I’m uploading a Shopify [REPORT TYPE: e.g., Sales by Product, Customer Cohort, Traffic and Conversion] report for my business. Here’s what I sell: [BRIEF DESCRIPTION: e.g., sustainable kids’ clothing, sizes 2–12].
The business question I’m trying to answer: [YOUR SPECIFIC QUESTION: e.g., Which products have the highest return rate? Which traffic sources convert best? What does repeat purchase behavior look like for customers acquired in Q4?]
Please analyze the data and give me:
- A direct answer to my question
- Any patterns or anomalies worth noting
- What this data suggests I should do or investigate further
[ATTACH SHOPIFY REPORT CSV OR PASTE DATA]
Variables to fill in:
- [REPORT TYPE]: the Shopify report you downloaded
- [BRIEF DESCRIPTION]: what your business sells
- [YOUR SPECIFIC QUESTION]: the decision you’re trying to make
What to expect: A direct answer plus 2–3 leads worth following up on. Add context the data can’t capture (out-of-stock periods, press mentions, seasonal factors) so the AI doesn’t misread anomalies.
Follow-up prompt
[CONTEXTUAL NOTE: e.g., “Traffic spiked in November because we got a feature in a major parenting publication.” / “We were out of stock on our top SKU for most of March.”] Does that change your interpretation? What should I look at next?
Putting it together
These three workflows compound on each other. Customer call analysis surfaces what your customers say they care about. Survey analysis shows you what a broader group of customers says at scale. Shopify data shows you what they actually do.
When the three tell the same story, you have high confidence.
When they diverge, for example, a customer says price doesn’t matter but conversion tanks when you raise prices. That gap is worth investigating. Run all three prompts regularly and feed the most interesting tensions back into the conversation: “My calls suggest X, but my Shopify data suggests Y. What might explain that?”
Tips for better results
- Cross-reference tools. Sachse runs call and survey analysis in at least two AI tools and compares outputs. Where they agree, act. Where they diverge, dig deeper.
- Lead with your business question. Before uploading any data, state the decision you’re trying to make. It focuses the output and cuts generic analysis.
- Check for missing coverage. AI will gravitate toward patterns it finds most legible. Scan the raw data yourself and explicitly ask about anything it didn’t address.
- Add context the data can’t see. External factors — seasonality, press, supply issues — don’t show up in a CSV. Naming them in the conversation produces much more accurate analysis.
- Iterate on the same conversation. Don’t start fresh for every follow-up question. Keep the context alive in one thread and build on each answer.
Ready for the full story?
Read The AI research stack behind a solo founder’s Nordstrom deal, featuring Marianna Sachse of Jackalo, published on the AI Lab by ActiveCampaign.
Prefer a step-by-step guide?
Get the step-by-step checklist that walks you through how to replicate Marianna’s customer data analysis workflow.
Related
More data from the AI Lab.