Handling Support Tickets with AI Without Burning the Planet
summary
How to handle support tickets with local, resource-light LLMs instead of huge cloud models like ChatGPT or GPT-5. The presenters describe a stepwise pipeline—ticket ingestion, thematic classification, similarity search against the global corpus, RAG to prepare answers, and knowledge-base capitalization—using Tiny Language Models (TLMs) and show a proof-of-concept with live demos, plus discussion of benefits, constraints and limits.
AI-written summary · about
chapters (6)
- 0:00AI and boilers intro
- 12:17AI models' environmental impact
- 21:24RAG optimized responses
- 34:21Autonomous agents local use
- 40:59Local AI ecological impact
- 45:13Q&A
related talks
47mOperationalization of AI/ML models and use cases for ChatGPT Q&A
32mAutomate customer ticket handling: from an idea to a production ready model
01h 04mBuilding Expertise Into Generated Conversations
21mChatGPT, LLMs, and LangChains - A Beginner's Guide
42mAdvancements in Evaluating Large Language Model Applications
36mProduction-Ready AI Agents: From LLMs to Small Language Models
43mAI Has Entered The Chat: Humans, AI, and the Art of Collaboration
51mA ChatGPT on your own data: how does it work?