> ## Documentation Index > Fetch the complete documentation index at: https://docs.tryprofound.com/llms.txt > Use this file to discover all available pages before exploring further. # Article Research Report > Produce a structured, multi-source research report with citations for a topic The **Article Research Report** node performs deep, multi-source research and produces a structured research report with citations, insights, and synthesized key findings for your chosen topic. This report is designed to serve as **reference material**—not an article itself—supporting strategists, writers, and downstream content-generation Agent nodes with authoritative, verifiable information. Unlike the [Research Snippets Generator](/agents/nodes/research/research-snippets-generator)(which extracts factual snippets), the Article Research Report agent creates a **full narrative research summary**, typically 800–1300 words, complete with inline citations and a reference list. Check out [Getting started with Agents](/agents/getting-started) to learn how to add this node to an Agent. ## When to use this node Use the Article Research Report when you need: * A comprehensive research foundation before writing a content brief or article * A narrative-format summary of key findings across many sources * A detailed explanation of trends, themes, frameworks, or definitions * A fact-rich report used to educate subject-matter experts or content teams * AEO-aligned supporting material to ground downstream content generation This node is ideal for workflows where accuracy, context, and depth matter. ## Node configuration Selecting the Article Research Report node opens its configuration panel on the right side of the Agent builder. ### Article Title (required) The article title or topic the research should focus on, for example: * “Emerging trends in healthcare technology” * “How AI is transforming supply chain management” * “Cybersecurity risks in financial services” This title guides the search queries and synthesis. ### Audience Segment The intended audience for the research output. This influences tone and depth: * Executive summaries * Technical explanations * Practitioner-level insights ### Company Name Used to ensure report framing aligns loosely with the brand context or vertical. ### Content Type Defaults to **General**, but can influence which explanations or structures are emphasized (e.g., guides vs. comparisons vs. thought leadership). ### Target Prompt If your article must support a specific user query used in AI systems, enter that here. For example: “What is predictive maintenance and how does it work?” This helps your Agent prioritize research aligned to key user intents. ### Output Label Assign a descriptive variable name for downstream referencing. Examples: * `research_report` * `background_research` * `topic_report` ## How the node works behind the scenes The Article Research Report node executes a multi-stage research and synthesis pipeline. The internal workflow includes: ### 1. Input evaluation and research setup The workflow prepares your inputs (title, content type, audience, company context) and determines which research patterns to use. For example: * “How-to” topics may require step-by-step explanations * Comparative topics may need entity lists * Thought-leadership topics require key themes and debates ### 2. Multi-engine topic research The node runs several rounds of focused research across tools like: * Tavily * Perplexity * Specialized LLM-based research routines Multiple queries are generated from your article title to capture: * Definitions * Trends * Statistics * Case studies * Expert opinions * Historical context * Practical frameworks This produces a large pool of raw research artifacts. ### 3. Deduplication and relevance filtering The workflow discards: * Low-authority sources * Irrelevant tangents * Redundant content * Outdated material It retains only high-quality, authoritative findings. ### 4. Structured evidence formatting Research results are converted into a structured JSON dataset that contains: * Summaries * Key facts * Source URLs * Extracted insights This dataset becomes the input to the synthesis model. ### 5. Research synthesis into a narrative report An LLM configured as a **research synthesizer** composes an 800–1300 word report following strict requirements: * Clear paragraph writing * Inline numbered citations: \[1], \[2], \[3], … * A reference list at the end * A content-type–specific structure: * **How-to** → step-by-step * **Comparative** → list of entities to compare * **Thought leadership** → themes and debates * **Opinion** → insights backed by evidence The synthesizer is instructed to: * Be factual * Avoid hallucinations * Include only information supported by cited sources * Be neutral and research-driven * Produce *only* the research report (no article draft, no commentary) ### 6. Final formatting The output is standardized into a structured, ready-to-use text block containing: * Title * Corresponding narrative content * Inline citations * References section This makes the report suitable for editors, strategists, or downstream LLM article generation. ## Output The final output is an **800–1300 word research report** containing: * A narrative overview of the topic * Key findings from multiple sources * Definitions, models, or frameworks * Trends, challenges, and future outlooks * Inline citations like \[1], \[2], \[3]… * A reference list at the end This report is intended to be a foundation for: * Content briefs * Articles * Thought-leadership pieces * Educational summaries * Editorial decision-making ## Example usage ### 1. Create a deeply researched content workflow 1. Article Research Report ​ 2. [Research Snippets Generator](/agents/nodes/research/research-snippets-generator) 3. [Create Content Brief](/agents/nodes/content-creation/create-content-brief) 4. [Generate Article](/agents/nodes/content-creation/generate-article) This ensures every step is backed by trustworthy research. ### 2. Editorial preparation Use this node to educate internal teams about a topic before writing. ### 3. Automated research pipelines Trigger recurring research reports on topics like: * Industry trends * Product categories * Seasonal themes * Competitor verticals ## Best practices * Include a **Target Prompt** if you want the report aligned to real user queries. * Provide a clear, specific **Article Title** to guide more accurate research. * Pair this agent with the [Research Snippets Generator](/agents/nodes/research/research-snippets-generator) for fact extraction. * Use the output label consistently for downstream LLM steps. * Treat this report as **research**, not as a final draft article.