Generative AI’s impact on software engineering team productivity
summary
Claim: StackSpot uses a hyper contextualization agent that pulls company Stacks, uploaded knowledge sources (21 items in the demo), project code and embeddings, then prompts an LLM to produce ready Kotlin Spring Boot code with OpenFeign clients and JUnit tests. Demo highlights: developers average 52 minutes coding daily (4h21/wk), McKinsey-style productivity uplift of 20–40%, automatic docs and line‑level explanations, COBOL business logic rewritten into modern service code for an equivalent interest rate calculation, and generated code inserted straight into the IDE and committed.
AI-written summary · about
chapters (7)
- 0:00The case for generative AI
- 3:46Hyper-contextualization
- 9:10Hands-on coding demo
- 19:17Modernizing legacy code
- 24:27Maintaining chat context
- 27:18System architecture overview
- 30:21Management and final call
related talks
42mSoftware Tomorrow
50mDeveloper Productivity – DIY (with LLM) or How I Learned to Stop Worrying and Love the AI
27mAI for software development: A reality check
27mSoftware engineering with GenAI
14mFull-Stack AI Engineering from AI-Native Coding to Knowledge, Models, Infrastructure & OSS
33mAI assisted software delivery
41mAI-Augmented Testing: How Generative AI, Prompt Eng Turn Testers Into Superheroes
14mDeveloper productivity in a post-AI world