Home · Business · Customer Service · Feedback & NPS analysis
As of 13 August 2026, AI can create an NPS report for your small business.
This still needs a person who signs their name to it.
Can you do it?
5 minutesto a draft.
30 minutesto something you’d act on.
Cost, all in£0
Skill neededchat-fluent
Who has to check ityou
What the alternative costsCustomGPT is a no-code custom chatbot from your business content with citations.
If this goes wrong: the report contains a wrong calculation or overconfident interpretation, which you can correct by comparing it with the source data before sharing it.
What to actually do
Hand it to a person
The route this page recommends
A person who owns the outcome does this end to end, worth it when the failure is dear.
Do it yourself
Second choiceA chat interface, chat-fluent skill, and roughly 30 minutes until you can act on the result.
How to actually do it
- Open the spreadsheet or survey platform export containing one row per response, remove names and contact details, and keep the NPS score, response identifier and any useful customer segment.
- Confirm the exact NPS question, survey period, response count and definitions of any segments, then write these in a short context note.
- Paste the context note and the cleaned response table into a chatbot with the prompt above, or upload the CSV if the chatbot supports file analysis.
- Ask the chatbot to produce the report and to flag duplicated, missing or ambiguous rows instead of filling them in.
- Open a spreadsheet and independently count scores from 0 to 6, 7 to 8 and 9 to 10, then calculate each percentage and subtract the detractor percentage from the promoter percentage.
- Compare the spreadsheet calculation, response count and segment totals with the drafted report, and trace each stated finding back to the supplied responses or comments.
- Remove any unsupported claims and anonymise quoted comments, then send the checked report to the people who need it for the relevant customer-service actions.
Prompt
Create an NPS report for my small business using the survey data and context below. Survey question: [paste the exact NPS question] Survey period: [insert period] Number of responses: [insert number, or calculate it only from the supplied rows] Customer segments to compare: [list segments, or say none] Business context: [paste relevant product, service, pricing or operational context] Survey data: [paste a table or upload a CSV with one row per response] Use the standard NPS method: scores of 9 and 10 are promoters, scores of 7 and 8 are passives, and scores of 0 to 6 are detractors. Calculate NPS as the percentage of promoters minus the percentage of detractors, using all valid responses. Do not invent missing responses, dates, segments, reasons or business context. State how many valid responses were used and show the promoter, passive and detractor counts and percentages. If the data is incomplete, duplicated or ambiguous, list the problem instead of guessing. Produce: 1. An executive summary in plain British English. 2. A calculation table that makes the NPS result reproducible. 3. Segment comparisons only where the segment is present in the data. 4. The main themes from any open-text comments, with representative comments labelled as excerpts rather than presented as statistically representative. 5. Three practical actions linked to the evidence, separating observed findings from recommendations. 6. A short limitations section covering sample size, response bias, missing data and any uncertainty. 7. A final list of every figure and claim that I should check against the original survey export before sharing the report. Do not claim that the result represents all customers unless the supplied data supports that conclusion. Do not expose names, email addresses or other unnecessary personal data in the report.
Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.
Use a tool built for this
The distant third
What it gets wrong
- AI cannot decide whether a change in NPS reflects genuine customer sentiment or a change in who answered the survey.
- AI cannot repair a biased sample or make a low response rate representative of your whole customer base.
- AI groups open-text comments by wording and may miss sarcasm, local knowledge or the commercial importance of an unusual complaint.
- AI cannot choose which customer-service action is worth the cost without your operational and financial judgement.
- AI cannot verify that the survey export is complete unless you compare it with the source system.
Even on a YES, the friction has a name: judgement under ambiguity and verification cost.
How we scored this
Five axes, each scored nought to two by hand: ten means AI carries the task cleanly, and the thresholds that turn a total into YES, PARTLY or NO are published in the methodology. Each axis name links to its definition.
| Axis | Score (0–2) |
|---|---|
| Output | 2 |
| Inputs | 2 |
| Verification | 2 |
| Liability | 2 |
| Effort delta | 2 |
| Total | 10 / 10 |
The methodology and its thresholds are published in full.
FAQ
- Can ChatGPT calculate my NPS?
- Yes. Give it the response scores and ask it to show the promoter, passive and detractor counts and percentages, then check the result in a spreadsheet before using it.
- What data do I need for an NPS report?
- You need the response scores and the exact survey question, plus the survey period and any customer segments you want to compare. Open-text comments and business context make the findings more useful, but the NPS calculation itself only needs valid scores.
- Can AI analyse the comments in an NPS survey?
- Yes. It can group comments into themes and draft a summary, but the themes still need checking against the original comments. Remove names, email addresses and other unnecessary personal data before uploading them.
- Can I trust an AI-generated NPS report?
- You can use it as a first draft, not as an unchecked source of truth. Recalculate the score, compare all totals with the survey export and remove conclusions that the responses do not support.
Nearby answers
- Can AI identify customers at risk of leaving from their feedback?PARTLY
- Can AI summarise my customer survey responses?YES
- Can AI analyse feedback from my Google reviews?YES
- Can AI benchmark my NPS score against similar UK businesses?PARTLY
- Can AI compare feedback from different customer segments?YES
- Can AI create a monthly summary of my customer feedback?YES
Assessed by gpt-5.6-luna (gpt-5.6-luna) on 2026-08-13, second-checked by an independent model. Wrong somewhere? Email [email protected] and it gets re-checked.
The newsletter
AI news, new answers and product picks, straight to your inbox.