Back to Research Notes Index
NBER / Stanford // 2023TOPIC: Automation
Generative AI at Work
AUTHORS: Erik Brynjolfsson, Danielle Li, Lindsay Raymond | REVIEWED: 2024-10-12
PAPER FINDING
NBER Working Paper 31161 field study showing generative AI tools provide the highest relative productivity boost to novice and mid-tier workers by codifying tacit knowledge.
MY INTERPRETATION
In my view, effective AI workflows act as operational force-multipliers rather than total human replacements when paired with human escalation controls.
MY IMPLEMENTATION / CONCEPTUAL APPLICATION
Conceptual application: I structured intent classification handlers in customer messaging workflows to route low-confidence edge cases to human supervisors.
EVIDENCE & METHODOLOGY
Analysis of NBER Working Paper 31161 empirical data.
RETRIEVAL_INTELLIGENCE // HYBRID_RAGVECTOR RETRIEVAL & RERANKING
STEP 01
Chunking
512 token splits
STEP 02
Embedding
Dense vectors
STEP 03
Qdrant Search
Cosine sim (k=25)
STEP 04
Cross-Encoder
Rerank top-5
STEP 05
Grounded Gen
With citations
SYSTEM LIMITATIONS & EDGE CASES
- • Productivity improvements depend on the currency and curation of underlying knowledge bases.