Bank evolution to data science-driven business model
summary
Banca Intesa Belgrade moves CRM from sales-force heuristics to production ML-driven campaigns, building bespoke binary classifiers in Python ML models for each product and using web behavior for a two-stage next-best-offer recommender (MBP1/MBP2). Feature set grows eightfold, models run in production with control groups, decision trees first then deep learning experiments, core banking swaps and COVID force repeated rebuilds, and the headline metric is a flagship personal-loan campaign delivering 7x conversion over random, making campaigns the dominant sales source.
AI-written summary · about
chapters (14)
- 0:00Intro
- 0:22Business Model
- 7:42Migration to Python
- 9:21Campaigns migrated
- 12:13Outbound campaigns
- 14:20Inbound sales
- 17:48Data sources
- 20:10Commercial information
- 25:05Quality of models
- 28:28Critical elements
- 30:58Critical activities
- 31:56Challenges
- 33:07Tradeoff
- 34:02Conclusion
related talks
29mAI-Driven Business Success: Measurable Results and Revolutionized Sales
57mAI has been achieved internally - The Journey
31mBuilding Data-Driven Client Relationship Management in Banking with Python
28mRecent approaches in retail banking marketing optimization
01h 03mAI Innovation for a Smarter Financial Ecosystem
27mData Science and ML with an emphasis on Banking...
37mApproaching AI in Banking
29mAI - Digital (r)evolution in banking