What this is worth, and how it lands

Lulu is starting from nothing on AI. The route below is deliberately staged: prove it in one country, widen to all seven systems, federate across six countries, then operate as one system with a standing governance cadence.

AED 372m captured to dateAED 766m modelled a yearFour phases

Captured against modelled

AED millions. Captured is agreed by finance; modelled is the calculated opportunity at current scope.

Group AI maturity

49%

Across ten verticals

Hypermarkets and stores

62%

Fresh food and production

55%

Supply chain and logistics

48%

E-commerce and delivery

58%

Merchandising, pricing and margin

51%

Customer, loyalty and marketing

64%

People and workforce

44%

Property, energy and facilities

32%

Finance, risk and shared services

38%

Private label and sourcing

35%

The route in

Each phase states what it proves, and what it needs from Lulu to start.

Weeks 1–8

Phase 1 · Prove

Scope: 3 systems, 9 stores, 1 country

Needs: Data access and one named owner per system

Proves: Availability, fresh waste, answered demand

Weeks 9–20

Phase 2 · Widen

Scope: 7 systems, 1 country, all stores

Needs: Agent memory in the lake, refusal thresholds agreed

Proves: Margin, loyalty activity, roster fill

Weeks 21–40

Phase 3 · Federate

Scope: 6 countries, 111 stores

Needs: Country owners and one operating cadence

Proves: Group margin, cost to serve, attrition

Continuous

Phase 4 · Operate

Scope: Whole group as one operating system

Needs: Standing governance and quarterly threshold review

Proves: Value captured against modelled

Value by vertical

Modelled annual value in AED millions, with the count of products behind it.