社内å•ã„åˆã‚ã›ã‚’AIエージェント化ã—ã¦çˆ†é€Ÿã§è§£æ±ºã§ãるよã†ã«ã—ãŸã¯ã˜ã‚ã«ï¼šç¤¾å†…å•ã„åˆã‚ã›ãŒã˜ã‚ã˜ã‚開発を削ã£ã¦ã„ã ダイニーã§ã¯ã€é£²é£Ÿåº—舗ã‹ã‚‰æ—¥ã€…ã€é‹ç”¨ã‚µãƒãƒ¼ãƒˆãƒãƒ¼ãƒ ã«å•ã„åˆã‚ã›ãŒå¯„ã›ã‚‰ã‚Œã¾ã™ã€‚ãã—ã¦ã€é‹ç”¨ã‚µãƒãƒ¼ãƒˆãƒãƒ¼ãƒ ã ã‘ã§ã¯è§£æ±ºã§ããªã„技術的ãªå•ã„åˆã‚ã›ã‚’開発ãƒãƒ¼ãƒ ã¸ã¤ãªã"å…¥å£"㌠社内å•ã„åˆã‚ã›ã‚·ã‚¹ãƒ†ãƒ ã§ã™ã€‚ 「会計ãŒã†ã¾ãã„ã‹ãªã„ã€ã€Œè¨å®šç”»é¢ã®æ“作ãŒåˆ†ã‹ã‚‰ãªã„ã€ã¨ã„ã£ãŸå£°ãŒã€ç¤¾å†…å•ã„åˆã‚ã›ã‚·ã‚¹ãƒ†ãƒ を通ã£ã¦æ—¥ã€…エスカレーションã•ã‚Œã¾ã™ã€‚ ãŸã ã€é‹ç”¨ãŒè»Œé“ã«ä¹—ã‚‹ã«ã¤ã‚Œã¦æ¬¡ã®èª²é¡ŒãŒé¡•åœ¨åŒ–ã—ã¾ã—ãŸã€‚ 1æ—¥ã‚ãŸã‚Šç´„8件ã®å•ã„åˆã‚ã›ãŒç™ºç”Ÿã™ã‚‹ 1件ã®èª¿æŸ»ã«æ•°æ™‚é–“ã‹ã‹ã‚‹ リードタイムã®ä¸å¤®å€¤ãŒ10æ—¥ ステータス管ç†ãŒå¼±ãã€ã‚¯ãƒãƒ¼ã‚ºã•ã‚Œãšã«æ”¾ç½®ã•ã‚Œã‚‹ 社内ã®å•ã„åˆã‚ã›å¯¾å¿œã¯ã€ç›®ç«‹ã¡ã«ãã„ã®ã«ç¢ºå®Ÿã«é–‹ç™ºãƒãƒ¼ãƒ ã®æ™‚間を削りã¾ã™ã€‚ãã“ã§ã€ç¤¾å†…å•ã„åˆã‚ã›ã‚·ã‚¹ãƒ†ãƒ ã‚’ AI エージェント化ã—ã€å®šåž‹çš„ãªèª¿æŸ»ã¨ãƒªãƒžã‚¤ãƒ³ãƒ‰ã‚’自動化ã—ã¦ã€å¯¾å¿œã‚’ãƒãƒ¼ãƒ 全体ã§ã‚¹ã‚±ãƒ¼ãƒ«ã•ã›ã‚‹æ–¹é‡ã«
Claude Codeã«é•·æœŸè¨˜æ†¶ã‚’æŒãŸã›ãŸã‚‰ã€å£æ‰“ã¡ã®è³ªãŒå¤‰ã‚ã£ãŸ
LLMã‚’ã‚‚ã£ã¨è³¢ãã™ã‚‹ï¼šãƒŠãƒ¬ãƒƒã‚¸ã‚°ãƒ©ãƒ•å®Ÿè·µå…¥é–€
How We Hacked McKinsey's AI Platform
https://x.com/i/status/2031416367610744960
GraphRAGを実際ã«æ§‹ç¯‰ã—ã¦åˆ†ã‹ã£ãŸã€Œä½¿ã†ã»ã©è³¢ããªã‚‹AIã€ã®ä»•çµ„ã¿
https://x.com/i/status/2029464311329099960
Improve Gen AI Search with Vertex AI Embeddings and Task Types | Google Cloud Blog
社内è¦ç¨‹RAGã®ç²¾åº¦ã‚’73.3% → 100%ã«æ”¹å–„ã—ãŸè©±
Agentic Workflows
Vertex AI Searchを使ã„ã“ãªã™å®Ÿè·µãƒ†ã‚¯ãƒ‹ãƒƒã‚¯ | GCTS 2026
Keyword search is all you need: Achieving RAG-level performance without vector databases using agentic tool use
Claude CodeãŒãƒ™ã‚¯ãƒˆãƒ«æ¤œç´¢ã‚’採用ã—ãªããªã£ãŸç†ç”±
Collections | xAI
https://x.com/i/status/2019724288774021215
AIエンジニアãŒLangChainを推奨ã—ãªã„ç†ç”±
State of AI | OpenRouter
https://x.com/i/status/2016365319204569351
完璧ãªä»•æ§˜ã¯ã€ã‚‚ã†è¨è¨ˆã§ããªã„ - Nothing ventured, nothing gained.
AIエージェントを「自己進化ã€ã•ã›ã‚‹ä»•çµ„ã¿
{{#tags}}- {{label}}
{{/tags}}