✨ $500 AI Visibility Audit — live at Spurlock Studios. Book the audit

Operations & Systems

🚀Personal Deep Research | Subflow

This n8n workflow conducts AI-driven deep research by iteratively generating search queries, scraping web content for learnings, and refining the research based on accumulated insights. It then compiles a comprehensive report from all gathered data, converting it into Notion…

Built · ~7.5 hours saved per week

Personal Deep Research | Subflow Overview Automation Goal This workflow automates the entire process of conducting in-depth, iterative online research on a given topic. It starts with a single query, recursively expands the research with AI-generated sub-queries, scrapes and analyzes web content, and culminates in a comprehensive, fully-formatted report delivered directly to a Notion page. The primary value is in transforming a multi-hour or multi-day manual research task into a fully autonomous process, delivering synthesized insights and cited sources. Technologies Used - n8n: The core orchestration platform, managing the complex recursive logic and data flow. - AI (via OpenRouter): Integrates multiple Large Language Models (LLMs) like OpenAI, DeepSeek, and Anthropic for various reasoning, generation, and formatting tasks. - Apify: A web scraping and automation platform used to perform web searches and extract content from websites. - Notion: Serves as both the user interface for initiating requests and the final destination for the generated research report. How the AI Integrates This workflow uses AI in four distinct and powerful ways to achieve autonomous research: 1. AI-Powered Query Expansion: Initially, an AI model analyzes the user's primary query to generate a set of more specific, targeted search engine queries. This ensures the research has sufficient breadth and explores various facets of the topic, mimicking how an expert human researcher would break down a problem. 2. AI Information Synthesis: After web content is scraped for each sub-query, an AI model reads through the raw text (up to 25,000 characters per source) and extracts the most crucial pieces of information, referred to as "learnings." This condenses vast amounts of text into dense, actionable insights. 3. AI Report Generation: Once the recursive research is complete, all collected "learnings" are fed into a final AI prompt. The model is instructed to act as an expert researcher and write a detailed, multi-page report in Markdown, synthesizing all the information into a coherent narrative. 4.

Tags: Web Scraping, AI Automation, Google, Notion, Research

Integrations: Notion, HTTP, Sub-workflow, Code

Related automations

Back to the Automation Library