AI agent team guide
AI Research Report Writing: Sources to a Reviewable Draft
Turn an authorized research dossier into a coherent, source-checked report with an accountable publication handoff.

Use an AI research report writing team to turn an authorized dossier into an argument a reader can follow and verify. This workflow covers outline, evidence, continuity and editorial handoff. It does not invent a study, turn repeated claims into independent evidence or publish a report without approval.
Build your four-agent team
| Agent template | Its job on this team | Give it | Reviewable return | Team |
|---|---|---|---|---|
Longform EditorLead agent | Build the report’s argument and preserve continuity. | Accepted question, audience and source pack. | A coherent draft with a visible claim structure. | Add to team |
Research Reading CuratorSupporting agent | Check source relevance, provenance and contradictions. | Permitted research dossier and material claim list. | A claim-source map with gaps and limitations. | Add to team |
Brand Voice EditorSupporting agent | Edit the accepted draft in the intended voice. | Approved argument and writing examples. | Clearer prose without strengthening unsupported claims. | Add to team |
Content ManagerSupporting agent | Prepare the publication handoff and next actions. | Accepted draft, assets and review requirements. | An owner-assigned editorial readiness checklist. | Add to team |
Set up your team in Finta
- Add each coworker individually. Use each Add to team action, review the import setup and complete that template’s Getting Started Skill. You can also open its marketplace profile to review its instructions and requirements first.
- Supply role-specific context. Give the lead the shared objective. Give each supporting agent only the authorized evidence and accepted prior return it needs. Importing the templates does not automatically connect outside systems or share all context.
- Review each handoff. Accept or correct the return before passing its evidence and open questions to the next role. Ask Aurora to help coordinate the work while you retain the decisions.
- Start with one manual pass. Review any supported routine separately before enabling recurrence. External messages, record changes and commitments need their own authorization.
Prepare your inputs
Supply the audience, decision or question, authorized source pack, publication format and approved product facts. Distinguish your original research from third-party material. Include source dates and limitations. State the rights and confidentiality boundaries for any material that should not appear in public.
Best practices worth bringing into the work
Purdue OWL’s outline guidance starts with purpose, audience and a logically ordered structure. Its argument guidance connects claims with evidence and meaningful counterarguments. Apply those habits before expanding a dossier into a long draft.
Run the workflow
1. Write a claim-led outline
Longform Editor states the central answer and the job of each section. Every section should advance the argument, explain evidence or resolve a reader question. Remove sections that merely repeat the introduction at greater length.
2. Check the evidence before drafting confidently
Research Reading Curator maps material claims to exact source locations and flags contradictions. Preserve the source’s population, period and limits. A report may say the available evidence is inconclusive; it must not invent consensus or original findings.
3. Draft, then edit the joins
Longform Editor drafts from the accepted outline. Check whether terms, metric definitions and examples stay consistent across sections. Brand Voice Editor improves clarity while preserving qualifications and evidence. Do not make a claim stronger simply because the stronger sentence reads better.
4. Prepare a human-owned publication packet
Content Manager checks citations, permissions, images, metadata, internal links and the reviewer list. Keep unresolved items visible. The publisher approves the actual final version and destination; importing the team does not connect or activate a CMS publishing routine.
Argument, claim-source and continuity map
Use this original worksheet with your own authorized evidence. The entries below are prompts for a review, not completed work.
| Section / reader job | Material claim | Evidence / limitation | Editorial action |
|---|---|---|---|
| Opening answer | What the report can support. | Strongest relevant sources and scope. | State the answer without exaggeration. |
| Analysis | Why the evidence supports the inference. | Exact locations and counterevidence. | Separate source findings from interpretation. |
| Conclusion | What the reader should do next. | Conditions and unresolved questions. | Name a useful action within the evidence. |
Worked example
Fictional example: a report dossier contains three articles repeating one underlying survey. The team cites the survey once, notes its population and removes the phrase “multiple independent studies.” The report becomes more useful by explaining the evidence accurately, not by increasing the number of citations beside an inflated claim.
Copy the team brief
Draft a report for [audience] answering [question] from the authorized dossier. First return a claim-led outline and source map. Preserve dates, definitions and limitations; identify repeated underlying evidence and counterarguments. After approval, draft for continuity and clarity. Prepare a publication handoff, not a live publish action.Review the return
Finish with a readable draft, a traceable claim map, consistent definitions and an accountable publication checklist. Mark fictional examples beside the example and label third-party research accurately. The next owner should know exactly which version is ready and what remains unresolved.
Before you start
Can an AI-written report be called original research?
Only if it genuinely reports original research with an accountable method and evidence. Synthesizing published sources can be valuable, but it should be described as synthesis or analysis rather than a new study.