#!/usr/bin/env python3 """ AI 使用场景每日收集任务 """ import sqlite3 import json import hashlib from datetime import datetime, timedelta from pathlib import Path # 数据库路径 DB_PATH = Path("/root/.openclaw/workspace/projects/ai_usecase_collector/data/ai_usecases.db") # 搜索结果（从 Tavily 获取） SEARCH_RESULTS = [ { "title": "Beyond the Agentic AI Hype: How Leading Enterprises Make It Actually Work", "url": "https://techcrunch.com/sponsor/shelf/beyond-the-agentic-ai-hype-how-leading-enterprises-make-it-actually-work/", "description": "In 2026, leading organizations are using AI agents to transform customer experiences, streamline operations, and execute work in new ways, driving real business impact. AI agents are applied across front-end customer experiences and back-office operations.", "source": "TechCrunch", "categories": "Enterprise, Customer Experience", "tags": "Agentic AI, Enterprise AI, Automation", "key_technologies": "AI Agents, LLM" }, { "title": "Nvidia shows future of self-driving cars, robots, AI agents", "url": "https://thenightly.com.au/business/nvidias-result-foreshadows-a-future-of-self-driving-cars-robots-and-personal-ai-agents--c-21762456", "description": "Personal AI agents that can list investment properties with right dimensions, arrange inspections, search for cheapest mortgages, and remind about upcoming birthdays.", "source": "The Nightly", "categories": "Personal Assistant, Real Estate", "tags": "Personal AI, Nvidia, Autonomous", "key_technologies": "AI Agents, Robotics" }, { "title": "7 Ways AI Will Transform US Manufacturing by 2030", "url": "https://www.industryweek.com/sponsored/article/55358765-7-ways-ai-will-transform-us-manufacturing-by-2030", "description": "Plant managers will become agentic orchestrators—managing fleets of purpose-built AI agents designed to deliver specific outcomes. Each agent handles predictive maintenance, quality inspection, supply chain disruptions, or production scheduling.", "source": "IndustryWeek", "categories": "Manufacturing, Automation", "tags": "Agentic Orchestration, Manufacturing AI", "key_technologies": "AI Agents, Digital Twins" }, { "title": "Four Ways Operators Are Using AI to Increase Competitiveness", "url": "https://www.telecoms.com/ai/four-ways-operators-are-using-ai-to-increase-competitiveness", "description": "Use of MCP (Model Context Protocol) to Integrate Systems and AI Agents. As Agentic AI expands across operators, MCP is set to become a key enabler for integrating AI with external systems.", "source": "Telecoms.com", "categories": "Telecommunications", "tags": "MCP, Agentic AI, Integration", "key_technologies": "MCP, LLM, AI Agents" }, { "title": "What Is Agentic Automation, and Why Does It Matter?", "url": "https://www.moveworks.com/us/en/resources/blog/what-does-agentic-automation-mean", "description": "Agentic automation combines LLMs that understand human text, large action models (LAM) that decide actions, and generative AI. It's used for customer service, HR operations, and IT support automation.", "source": "Moveworks", "categories": "Enterprise Automation", "tags": "Agentic Automation, LAM, LLM", "key_technologies": "LLM, LAM, AI Agents" }, { "title": "Agentic AI: How It Works and 7 Real-World Use Cases", "url": "https://www.exabeam.com/explainers/ai-cyber-security/agentic-ai-how-it-works-and-7-real-world-use-cases/", "description": "Agentic AI vs AI Agents: AI agents follow predefined rules while agentic AI can reason, plan, and adapt. Real-world use cases include cybersecurity, fraud detection, and automated threat response.", "source": "Exabeam", "categories": "Cybersecurity", "tags": "Agentic AI, Cybersecurity, Threat Detection", "key_technologies": "AI Agents, Machine Learning" }, { "title": "Building Autonomous Systems: A Guide to Agentic AI Workflows", "url": "https://www.digitalocean.com/community/conceptual-articles/build-autonomous-systems-agentic-ai", "description": "Agentic workflows perform interpretation and execute plans based on high-level objectives. Can recursively decompose tasks into subtasks. Agents work together, engage in negotiation, and transfer tasks.", "source": "DigitalOcean", "categories": "Workflow Automation", "tags": "Autonomous Systems, Multi-Agent", "key_technologies": "Multi-Agent Systems, LLM" }, { "title": "The Complete Guide to Building Agentic AI Workflows That Scale", "url": "https://galileo.ai/blog/ai-agentic-workflows", "description": "Customer service enhancement: Agents learn from every interaction, trimming average handle times. HR operations automation: Time-off approvals, employee onboarding, and policy Q&A.", "source": "Galileo AI", "categories": "Customer Service, HR", "tags": "Scalable Workflows, Customer Experience", "key_technologies": "AI Agents, LLM" }, { "title": "10 Agentic AI Examples and Use Cases", "url": "https://boomi.com/blog/10-agentic-ai-use-cases/", "description": "Integration and automation examples across various industries. AI agents handling data integration, business process automation, and event-driven workflows.", "source": "Boomi", "categories": "Integration, Business Process", "tags": "Integration, BPA, Event-Driven", "key_technologies": "AI Agents, Integration Platforms" }, { "title": "Agentic AI, explained", "url": "https://mitsloan.mit.edu/ideas-made-to-matter/agentic-ai-explained", "description": "JPMorgan Chase is exploring AI agents to detect fraud, provide customized financial advice, and automate loan approvals and legal document processing.", "source": "MIT Sloan", "categories": "Financial Services", "tags": "Banking, Fraud Detection, Automation", "key_technologies": "AI Agents, LLM" }, { "title": "Understanding AI Agentic Workflows", "url": "https://www.atlassian.com/blog/artificial-intelligence/ai-agentic-workflows", "description": "Traditional automation follows exact steps and breaks on surprises. Agentic AI workflows can handle unexpected situations, make decisions dynamically, and recover from errors.", "source": "Atlassian", "categories": "Workflow Design", "tags": "Adaptive Automation, Error Handling", "key_technologies": "AI Agents, Workflow Engines" }, { "title": "Future-Proofing Manufacturing Operations With AI", "url": "https://www.forbes.com/sites/forbes-listmaker/2026/03/03/future-proofing-manufacturing-operations-with-ai/", "description": "Modern AI systems can analyze data, plan, prioritize, simulate scenarios, and continuously improve. AI agents take concrete operational roles in manufacturing.", "source": "Forbes", "categories": "Manufacturing", "tags": "AI Operations, Simulation", "key_technologies": "AI Agents, Digital Twins" }, { "title": "AI Is The 21st Century Force Multiplier", "url": "https://www.forbes.com/sites/chuckbrooks/2026/02/27/ai-is-the-21st-century-force-multiplier/", "description": "Healthcare uses AI for precision diagnostics, medication discovery, and predictive medicine. Financial services use AI for fraud detection, risk modeling, and algorithmic trading.", "source": "Forbes", "categories": "Healthcare, Finance", "tags": "Force Multiplier, Diagnostics, Trading", "key_technologies": "AI, ML, LLM" } ] def get_db_connection(): """获取数据库连接""" conn = sqlite3.connect(DB_PATH) conn.row_factory = sqlite3.Row return conn def insert_scene(conn, scene): """插入场景到数据库""" # 使用 URL 生成唯一 ID scene_id = hashlib.md5(scene['url'].encode()).hexdigest()[:16] cursor = conn.cursor() # 检查是否已存在 cursor.execute("SELECT id FROM scenes WHERE original_url = ?", (scene['url'],)) existing = cursor.fetchone() if existing: print(f"✓ 已存在: {scene['title'][:50]}...") return False # 插入新记录 cursor.execute(""" INSERT INTO scenes ( id, title, description, original_url, source_platform, categories, tags, key_technologies, innovation_score, is_new, discovery_date, last_updated ) VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?) """, ( scene_id, scene['title'], scene['description'], scene['url'], scene['source'], scene['categories'], scene['tags'], scene['key_technologies'], 7, # innovation_score 1, # is_new datetime.now().isoformat(), datetime.now().isoformat() )) conn.commit() print(f"+ 新增: {scene['title'][:50]}...") return True def get_recent_scenes(conn, days=7): """获取最近 N 天的新场景""" cutoff = (datetime.now() - timedelta(days=days)).isoformat() cursor = conn.cursor() cursor.execute(""" SELECT * FROM scenes WHERE discovery_date > ? ORDER BY discovery_date DESC """, (cutoff,)) return cursor.fetchall() def get_top_scenes_by_innovation(conn, limit=10): """获取创新分数最高的场景""" cursor = conn.cursor() cursor.execute(""" SELECT * FROM scenes ORDER BY innovation_score DESC, discovery_date DESC LIMIT ? """, (limit,)) return cursor.fetchall() def get_category_stats(conn): """获取分类统计""" cursor = conn.cursor() cursor.execute(""" SELECT categories, COUNT(*) as count FROM scenes GROUP BY categories ORDER BY count DESC """) return cursor.fetchall() def main(): """主函数""" print("=" * 60) print("AI 使用场景每日收集任务") print("=" * 60) conn = get_db_connection() # 插入搜索结果 print("\n📥 收集新场景...") new_count = 0 for scene in SEARCH_RESULTS: if insert_scene(conn, scene): new_count += 1 print(f"\n✅ 新增 {new_count} 个场景") # 获取统计信息 print("\n📊 数据库统计:") cursor = conn.cursor() cursor.execute("SELECT COUNT(*) as total FROM scenes") total = cursor.fetchone()['total'] print(f" 总场景数: {total}") cursor.execute("SELECT COUNT(*) as new FROM scenes WHERE is_new = 1") new_total = cursor.fetchone()['new'] print(f" 新场景数: {new_total}") conn.close() return { 'new_count': new_count, 'total_count': total } if __name__ == "__main__": stats = main() print(f"\n📋 任务完成! 新增: {stats['new_count']}, 总计: {stats['total_count']}")