Home · Business · Customer Service · Feedback & NPS analysis

PARTLY

As of 13 August 2026, AI can only partly benchmark your NPS score against similar UK businesses.

This still needs a person who signs their name to it.

Can you do it?

15 minutesto a draft.

n/ait cannot be self-verified.

Cost, all in£0

Skill neededpower-user

Who has to check ita colleague

What the alternative costsThe supplied tool data gives no price for a human NPS benchmarking service.

If this goes wrong, you may treat an incomparable peer benchmark as evidence and direct customer-service investment towards the wrong problem.

What to actually do

  1. 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.

  2. Use a tool built for this

    Second choice
  3. Do it yourself

    The distant third

    A chat interface gets you a draft, but you cannot verify it yourself. That is the catch.

    How to actually do it

    1. Export your NPS survey results from the survey or CRM system and record the exact question wording, scale, survey period, response count, promoter count, passive count and detractor count.
    2. Gather current UK peer benchmark reports or published datasets and save the source URL, publication date, sector, customer type, geography, survey wording, sample size and response rate for each one.
    3. Open a spreadsheet and check that your promoter, passive and detractor counts add up to the total number of responses, then calculate the percentages and NPS independently.
    4. Paste the business context and the checked figures into the prompt, followed by the complete peer benchmark details and source links.
    5. Ask the chatbot to produce the comparison table and to exclude any benchmark whose methodology or population is not sufficiently comparable.
    6. Compare every reported figure with your spreadsheet and each cited source, then ask a colleague familiar with your customers and survey method to confirm that the remaining comparisons are fair before using the conclusion.

    Prompt

    Analyse my Net Promoter Score against the UK peer benchmarks supplied below. Use only the data I provide and do not invent companies, figures, sources, dates or industry averages.
    
    My business context:
    - Sector: [sector]
    - Business model: [B2B or B2C, and any relevant detail]
    - Customer type: [customer description]
    - Geographic scope: [UK-wide or region]
    - Survey period: [start date to end date]
    - Survey question and scale: [exact wording and scale]
    - Number of responses: [number]
    - Promoters: [number]
    - Passives: [number]
    - Detractors: [number]
    - My NPS, if already calculated: [score or leave blank]
    
    Peer benchmark data:
    [paste each source, its publication date, sector, business type, customer population, survey wording and scale, sample size, response rate if available, NPS and source URL]
    
    Tasks:
    1. Calculate my NPS as the percentage of promoters minus the percentage of detractors. Show the calculation and state how rounding was handled.
    2. Compare my score only with peer data that is reasonably comparable on sector, customer type, geography, survey wording, scale and survey period.
    3. Put every benchmark into a table with its source, date, sample size, comparability limits and difference from my score.
    4. Separate direct comparisons from weak or unsuitable comparisons.
    5. Do not claim that a difference is statistically significant unless the supplied data supports that conclusion. State what cannot be concluded from the data.
    6. Give three practical questions I should investigate next, without recommending a specific investment or claiming that NPS alone proves customer loyalty.
    7. End with a short plain-English conclusion labelled 'What this benchmark does and does not show'.

    Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

What it gets wrong

  • It cannot discover a representative peer group from your NPS score alone.
  • It cannot tell whether a published benchmark used comparable customers, wording, sampling or response methods unless those details are supplied.
  • It cannot turn a difference in NPS into a reliable explanation of why customers feel differently.
  • It can present an apparently precise ranking even when the underlying benchmark sources are weak or outdated.
  • It cannot take responsibility for customer-service or commercial decisions made from the comparison.

What caps this at PARTLY: real time truth, verification cost and context depth.

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.

AxisScore (0–2)
Output1
Inputs1
Verification0
Liability1
Effort delta1
Total4 / 10

FAQ

Can AI benchmark my NPS score?
Partly. AI can calculate your score and compare it with benchmark figures you provide, but it cannot create a trustworthy comparison group or verify that the sources are genuinely comparable.
What data do I need to benchmark my NPS?
Provide the exact survey question and scale, survey period, total responses, promoter, passive and detractor counts, customer type, sector and geography. For each peer benchmark, also provide the source, publication date, sample size, response rate and survey method where available.
Can ChatGPT compare my NPS with competitors?
It can compare your score with competitor or sector figures that you paste into the chat. It cannot reliably obtain private competitor data or decide that two scores are comparable without the underlying survey context.
How reliable is an AI NPS benchmark?
The arithmetic can be checked, but the benchmark is only as reliable as the peer data and methodology behind it. Check every source and have someone who understands your survey design assess whether the comparison is fair.

Nearby answers

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.