You don't need AI; You need to know what your data is.
summary
Claims that 93% of data science projects never reach production leads to a concrete remedy: stop fetishizing models and fix data. Uses Wardley maps to show innovation builds on commodities, gives a failed home‑insurance computer vision example where public image datasets miss South African brands, and argues for data catalog + data governance, de‑identification pipelines, and hiring data engineers rather than generic data scientists. Recommends reorganizing into interdisciplinary R&D/product teams and creating decision scientists focused on hypotheses and measurable business value.
AI-written summary · about
chapters (7)
- 0:00Introduction and welcome
- 3:18The AI frustration problem
- 12:29Wardley mapping and strategy
- 19:51Data context and commodification
- 24:44Fixing the data science mess
- 32:47Rethinking data science roles
- 45:51Q&A and closing remarks
related talks
26mThe Hidden Foundation: How Effective Data Management Determines AI System Success
20mWhy AI Projects Fail: The Missing Link of a Strong Data Strategy
09mUnlocking the value of your data
33mDataEngBytes 2023 - MEL-T1-02 - Kendra Vant
17mFocus on Data, on the Road to AI
32mLeading with Data: Crafting an Effective Strategy for 2023 and Beyond
41mGetting Your Data AI-Ready
26mUnlocking the Power of Data-Centric AI.