A BETTER WAY TO SHARE THE WORK
Bring research together without flattening the differences
The useful pattern is not always the one every source appears to agree on.
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 can an AI agent help synthesize research responsibly?
Give your agent an authorized source set and a specific question. Ask it to map observations to findings while retaining each source's method, audience and limitations. A reviewed HTML synthesis can show agreement, disagreement and next questions without pretending a selected bundle is a complete evidence base.
01 / SEE THE DIFFERENCE
Two sources, two different moments of confusion
Fictional non-sensitive usability notes, not population research.
Source A: five first-time volunteers tried a sample report; three asked what a label meant. Source B: eight selected support notes from existing users; five mention finding a saved report. Participants may overlap and this is unknown. Methods, tasks and audiences differ. Neither source is representative. Proposed interpretation: test terminology with newcomers and retrieval with existing users rather than claim one universal main problem. No combined rate or approved redesign.
- Source A
- Five first-time volunteers used a sample report; three asked about a label.
- Source B
- Eight selected existing-user support notes; five mention finding a saved report.
- Comparison limit
- Methods, tasks and audiences differ; participant overlap is unknown and neither source is representative. No combined rate.
- Next question
- Proposed: test newcomer terminology and existing-user retrieval separately. No redesign is approved.
Open a complete fictional HTML example, with its source notes →
Create my research synthesis02 / MAKE IT YOURSELF
Start with the evidence. Shape the page around the reader.
Define the boundary of the evidence
State which sources were supplied and why they are relevant. Distinguish a synthesis of selected notes from a systematic review. Do not imply your agent searched every study or accessed research it was not given permission to use.
Keep observations connected to their sources
Give each claim a source identifier and retain the method and context needed to interpret it. Repeated mentions can come from the same participant or event; do not count them as independent evidence without checking.
Explain differences before combining results
Different tasks, samples and measurements can produce apparently conflicting findings. Put those differences beside the comparison. Do not pool percentages, average incompatible scales or turn qualitative notes into prevalence estimates.
Separate what you learned from what to do next
Use HTML to show the finding, supporting and conflicting evidence, limitation and next question. Label any proposed action as interpretation, not a study result. Review privacy and every source link before sharing; keep raw participant data in the authorized research system.
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.
Synthesize only the authorized research sources I provide around [question]. Ask for missing source context or scope. State the source-set boundary. Map each observation and finding to a source identifier, method, audience and limitation. Preserve disagreement, unknown participant overlap and missing evidence. Do not pool incompatible counts, invent prevalence, claim a systematic review or stronger causal conclusion, or access unapproved material. Separate source findings from your proposed interpretations and next questions. Exclude identifying participant data and use only approved excerpts. Create readable static HTML with no scripts or forms and my saved style if available. Show me the draft and source checks. Do not publish or change access without separate explicit approval.
04 / BEFORE YOU SEND
A polished page still needs your judgment.
- Is the selected source-set boundary stated?
- Can every finding be traced to its source and context?
- Are disagreements and incompatible measures preserved?
- Are proposed actions separate from evidence and raw participant information excluded?
When HTML fits
A cross-source reading page for a bounded product or service question, with evidence and limits visible beside the synthesis.
When another format fits better
Do not substitute this for a systematic review or qualified medical, legal or other high-stakes evidence assessment. Keep raw research, consent records and analysis tools in their authorized systems.
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
How is this different from explaining one paper?
A paper explainer preserves one study's argument and boundaries. A synthesis compares a selected set and must also explain source selection, conflicting evidence and differences in methods.
Can I count mentions across interviews?
Only with a verified definition and deduplication method appropriate to your data. A mention count is not automatically a count of people or a population percentage.
What if sources disagree?
Show the disagreement and relevant context. It may identify separate audiences or a question to investigate; do not hide it to produce a confident headline.