Prosus' LCM-Promo: LLM uplift model cuts promo cost 6% in A/B test
ProsusAI · hf · 2026-11-10
Prosus and iFood researchers published LCM-Promo, an LLM fine-tuned on transactional commerce data that predicts promotion response to guide allocation.
- The model learns monotonic promotion-response curves from production logs, interpreting single-token outputs as conversion-threshold probabilities conditioned on session context like delivery-day shifts
- In a production A/B test on churn reactivation, it improved conversion by 1% and cut promotion cost by 6% versus the Segment-EMA baseline
- Demonstrates a path for task-specific fine-tuned LLMs in commercial operations rather than general chat
More from Venture
- Gary Marcus flags Goldman pushing Anthropic IPO while pulling its tools internally — GaryMarcus · 2026-11-11
- Nat Lambert: human expertise to diffuse AI's benefits is vastly underestimated — natolambert · 2026-11-11
- YC F26 Startup Elimentary Launches AI Agents for Private Credit Ops, Cuts NAV Checks to 30 Minutes — ycombinator · 2026-11-11
- Anthropic sets IPO trading debut for Thursday, November 19 — firstadopter · 2026-11-11
- Profound founder tried 12-15 ideas before hitting one that took three days — soleio · 2026-11-11
- AMD hits all-time high as traders hedge NVIDIA risk in AI buildout — BenBajarin · 2026-11-11