In this episode of the Databricks Diaries, Sunando Das, Global head of Predictive Analytics at Unilever shows how they approach AI readiness at scale, balancing proven predictive models with emerging generative AI to drive measurable business growth. Drawing from years of applied AI across demand forecasting, marketing, innovation, and investment efficiency, Sunando estimates AI's potential to unlock 2–3% incremental growth when applied with discipline and focus.

The conversation tackles one of the biggest challenges in AI today: why only around 5% of generative AI pilots make it into production. Sunando explains why closing this gap requires more than better models, it demands cultural change, shared ownership, and a clear expectation that experiments must convert into real business outcomes.

We explore how cross-functional collaboration between business experts and technical teams is essential, why AI must be treated as a shared business asset rather than a data team initiative, and how conversational interfaces and agent-based workflows are reshaping how organisations interact with data. From continuous upskilling to test-and-learn experimentation, this episode offers practical guidance for organisations looking to move beyond pilots and turn AI into sustained impact.