A BETTER WAY TO SHARE THE WORK
Share survey findings without overstating what people said
Make the responses useful without pretending your respondents speak for everyone.
New to creating with an agent? The request will guide it through the task. Bring your notes; you don’t need to write HTML.
How do I turn survey results into a useful page with my agent?
Supply de-identified results, the exact questions, recruitment method, and reviewed counts. Ask your agent to explain each finding with its correct denominator, show missing answers and overlapping themes, and keep sample limitations beside the claims. Review the analysis before sharing a static HTML summary with only audience-safe evidence.
01 / SEE THE DIFFERENCE
Twenty of thirty-two—not most customers
A fictional voluntary navigation survey. No population-wide estimate is supported.
Forty people submitted a voluntary survey from an in-product invitation. Thirty-two answered 'How easy was it to find the guide?'; 20 chose hard to find and 12 chose easy to find. Eight skipped that question. The eligible invitation count is unknown. No weighting was applied. Open text from 15 respondents was coded into overlapping themes: navigation 9, terminology 7; one response can have both. No quotation text is approved. Propose five follow-up conversations, not a confirmed redesign.
- Question result
- 20 of 32 answering respondents chose hard to find; 12 chose easy to find.
- Missing answers
- Eight of forty respondents skipped that question.
- Sample boundary
- Voluntary in-product invitation; eligible invitation count unknown and no weighting applied.
- Themes
- Among 15 open-text respondents, navigation appears in 9 and terminology in 7. Themes overlap.
- Next question
- Propose five follow-up conversations. The survey does not approve a redesign.
Open a complete fictional HTML example, with its source notes →
Create my survey findings report02 / MAKE IT YOURSELF
Start with the evidence. Shape the page around the reader.
Describe who actually answered
State how respondents were recruited and who was eligible. Voluntary replies do not automatically represent all customers. Preserve known nonresponse and sampling limitations; do not invent weighting or a margin of error for a sample that does not support it.
Use the denominator for each question
A survey with forty respondents may have only thirty-two answers to a particular question. Show that count beside the result. Distinguish skipped questions, excluded responses, and genuine zero values. Verify calculations outside the HTML page before writing the headline.
Explain themes without manufacturing consensus
Use an explicit coding method and retain contradictory or less common views when they matter. Multiple themes can occur in one response, so theme totals may exceed the respondent count. Do not invent quotations or treat selected anecdotes as frequency evidence.
Connect the finding to a bounded next step
Explain what the responses suggest investigating, not what they prove about every user. Keep proposed changes separate from research findings. Remove identifiers and sensitive free text, then have an authorized reviewer check the complete summary and sharing audience.
03 / TRY THIS PROMPT
Copy this into your agent. Add your source material.
Use the conversation where you already did the work, or start one with your notes. Replace the placeholders and supply any missing context; only attach information you are allowed to share with your agent.
Get the ready-to-copy creation request for guided questions, this task, and a draft review before anything is published.
Create a survey findings page using only authorized de-identified responses, exact questions, sampling details, and reviewed counts. Ask for missing essentials: recruitment, field dates, question wording, denominators, missing answers, exclusions, and coding method. Show counts beside percentages and preserve question-specific denominators. Keep respondent findings distinct from population claims. Do not invent representativeness, weighting, margin of error, causal conclusions, quotations, or missing counts. Explain overlapping themes and keep observed findings separate from proposed actions. Exclude identifiers and sensitive free text; use only explicitly approved excerpts. Create static HTML with no scripts, forms, survey collection, or live-data claims. Use my saved artifact style if available. Show the draft, calculations, and uncertain claims for review. Do not publish or change access without separate explicit approval.
04 / BEFORE YOU SEND
A polished page still needs your judgment.
- Does each result identify who answered that question and how many skipped it?
- Are recruitment limits and any weighting accurately described?
- Are themes, frequencies, and quotations traceable to the supplied method and evidence?
- Are generalizations and proposed actions distinct from observed responses?
- Could free text or small-group details identify a respondent?
When HTML fits
A readable findings companion for an approved audience, connecting reviewed aggregate evidence to a practical discussion without exposing a response spreadsheet.
When another format fits better
Use survey and analysis tools for collection, response management, weighting, and statistical work. Seek appropriate research review for consequential inferences. A static page is not a survey form, analysis engine, or guarantee of anonymity.
FROM DRAFT TO A LINK
Make it clear. Make it yours. Then share it.
Keep creating and revising in the agent you already use. With a compatible connection, share/artifacts publishes the reviewed HTML so you do not have to rebuild it in another editor. Ask your agent to use your saved artifact style for headings, color, and tone; review the result before sending it.
Choose who should be able to open the page and verify the returned link. For corrections to the same deliverable, update the existing artifact without changing its link. Publish a separate artifact when the old edition needs to remain available.
These are static pages, not live apps: no scripts, forms, or automatic data refresh. Source collection and scheduling depend on your agent and its available tools, not on share/artifacts.
Make this with my agentQuestions before you start
Can I show percentages?
Yes, when the denominator and calculation are valid. Include counts and the question context so readers understand what the percentage describes. Do not use the total customer base when the result comes only from respondents.
Why can theme counts exceed the number of comments?
If the documented coding method allows multiple themes per response, one comment can contribute to several categories. Explain that overlap rather than forcing themes to sum to one hundred percent.
Should I include customer quotes?
Only supplied excerpts approved for this audience, with identifying details removed and meaning preserved. Quotes illustrate a perspective; they do not establish its prevalence.