Workflow
Customer Support
icon
Search documents
X @IcoBeast.eth🦇🔊
IcoBeast.eth🦇🔊· 2025-09-05 16:33
Market Opportunity - The first company to develop indistinguishable AI agents for customer support/community management will achieve unicorn valuation rapidly [1]
Cisco TAC’s GenAI Transformation: Building Enterprise Support Agents with LangSmith and LangGraph
LangChain· 2025-06-23 15:30
AI Customer Support Challenges & Solutions - Cisco faced challenges in scaling support for mass scale issues, with potentially tens of thousands of customers opening cases simultaneously [2] - Monitoring AI agent performance and driving improvements proved difficult [2] - Integrating multiple AI agents (up to six) using different tools was unreliable and time-consuming [8][9] Langchain & Langsmith Implementation - Langchain facilitated rapid prototyping by enabling easy model swapping [3] - Langsmith provided full visibility into AI applications, replacing custom observability solutions [4] - Langsmith connected subject matter experts (SMEs) with developers, enabling direct feedback on RAG retrieval [4][5] - Langraph allowed breaking down code into modular nodes, providing greater control over flows [10] - Langraph platform and Langchain Agent Protocol standardized agent communication, simplifying integration and improving scalability [11] Business Impact - Cisco enhanced customer experience through improved observability and faster development cycles [12] - Langsmith accelerated development and prototyping, identifying what was working well and what wasn't [9] - The firewall assistant was the first application, leveraging domain expertise to improve in-product assistance [6][7]