Case study / Delivery record

AI supply-chain suite on Google Cloud for global CPG and retail

As Associate Project Lead for Pluto7’s AI and ML portfolio, I governed forecasting and digital-twin programmes for global CPG and retail businesses. The work improved forecast accuracy by 22%, reduced stock-outs by 18% and lifted OTIF by 12%.

  • Colgate-Palmolive
  • Tory Burch
  • AB InBev
  • LIXIL
  • Tamer Group
  • Vertex AI
  • +22% forecast accuracy
  • −18% stock-outs
  • +12% OTIF

01

Context

Reactive planning constrained working capital and service levels.

Enterprise supply chains running on reactive planning, with forecasting accuracy the constraint on working capital and service levels.

02

Challenge

Predictive models had to become adopted operating products.

Move from reactive to predictive: demand forecasting, digital-twin simulation and inventory visibility, delivered as products clients would adopt, not as data-science experiments.

03

What I did

Business KPIs, data foundations and executive reporting moved together.

Led delivery of the forecasting suite and digital-twin platform on Vertex AI, BigQuery, Pub/Sub and Firestore; designed AI reporting agents that cut executive reporting effort 40%; redefined supply-chain KPIs with client leadership; drove USD 500K+ in expansions.

04

Outcome

Forecast accuracy translated into supply-chain reliability.

+22% forecast accuracy, −18% stock-outs, +12% OTIF, +14% service reliability.

05

What it taught me

Every AI programme starts as a data-quality programme.

Every AI project is a data-quality project wearing a different hat. If the catalogue and inventory data is a mess, the model industrialises the mess at speed.

Programme timeline
  1. 01

    KPI definition

  2. 02

    Data foundation

  3. 03

    Forecasting suite

  4. 04

    Digital twin

  5. 05

    Portfolio expansion

← All work