LLMs: Powering Lightning Fast Product Innovation
summary
Clarify uses GPT APIs to turn call transcripts into smart summaries and action items, built end-to-end in about 15 days by chunking long transcripts, calling GPT for per-chunk summaries, then re-summarizing into one output. Engineering notes: handle hallucinations via constrained prompts, reduce cost by batching requests, force JSON output to parse reliably, and iterate prompts ruthlessly. Deployed to users on production data (600M signals, 0.5B emails ingested historically) and receives usable feedback.
AI-written summary · about
chapters (9)
- 0:00Introduction to Clari and LLMs
- 3:30What are LLMs and their history
- 9:33Early adopters of LLMs
- 11:20Wingman and call summaries
- 13:29Engineering smart summaries
- 15:59Prompt engineering challenges
- 18:49Cost and output validation
- 22:25User impact and future plans
- 23:51Conclusion and next wave
related talks
45mHow to use GPT in your projects: A Case Study
24mHarnessing the Power of LLMs in Advanced Analytics
01h 15m10x Development: LLMs For the working Programmer
41mConnect GPT with your data: Retrieval-augmented Generation
01h 00mGoing beyond ChatGPT: introduction to prompt engineering & LLMs
39mHow ChatGPT helps to boost your development efficiency
01h 00mNext-Gen Software Architecture with Large Language Models (LLMs) like ChatGPT
29mLLM-Based Development