Lulu Hypermarket

Lulu Retail Holdings

AI Operating System

DeployOne
Viewing as

Operating scope

Country head

Capability

All capabilities

Active handovers

0

Outcomes written back

0

Information density is explain.

AI product portfolio

39 AI products across ten verticals

Each product is a working thing with an owner, the humans in the loop, the agents behind it, the feeds it reads, what it writes back, and the limit it may not cross. Nothing here is a slide.

AED 766m modelled a year19 live10 piloting

Products by stage

Modelled value by stage

AED millions a year.

Adoption of live products

Share of eligible stores or teams using it weekly.

Hypermarkets and stores

Shelf vision and gap agent

Live

Camera and till evidence reconciled into a shelf truth feed per aisle.

Modelled value

AED 41m a year

Adoption

78%

Owner:
Country retail director
Humans in the loop:
Section heads confirm each gap before a pick is raised.
Agents:
Shelf Sentinel, Replenishment Planner
Reads:
Aisle cameras, Till lines, Backroom stock, Planogram master
Writes:
Gap picks, Planogram breaks, Phantom stock flags

No stock write-off without a named human approval.

Open Hypermarkets and stores

Hypermarkets and stores

Product locator concierge

Live

Answers where an item sits, in aisle and bay, in the shopper's language.

Modelled value

AED 12m a year

Adoption

64%

Owner:
Store operations
Humans in the loop:
Customer service desk picks up anything the agent refuses.
Agents:
Aisle Concierge, Language Router
Reads:
Bay map, Live stock, Arabic and Malayalam phrase bank
Writes:
Aisle and bay answer, Substitute offer, Unserved demand log

Never promises stock it cannot see.

Open Hypermarkets and stores

Hypermarkets and stores

Queue and checkout forecaster

Piloting

Predicts till load 40 minutes ahead and asks for a named opener.

Modelled value

AED 10m a year

Adoption

31%

Owner:
Store manager
Humans in the loop:
Duty manager approves each till opening.
Agents:
Queue Forecaster, Roster Broker
Reads:
Footfall counters, Till states, Roster
Writes:
Open till request, Break reshuffle

Cannot move a colleague without their supervisor.

Open Hypermarkets and stores

Hypermarkets and stores

Planogram drift watcher

Building

Flags where the floor has drifted from the agreed plan.

Modelled value

AED 6m a year

Adoption

12%

Owner:
Category head
Humans in the loop:
Space planners sign off every reset.
Agents:
Shelf Sentinel
Reads:
Aisle cameras, Space plan
Writes:
Reset task, Drift trend

Photographic evidence attached to every claim.

Open Hypermarkets and stores

Fresh food and production

Countdown pricing agent

Live

Times markdowns by hour, shelf life and footfall instead of a fixed rule.

Modelled value

AED 29m a year

Adoption

71%

Owner:
Fresh category head
Humans in the loop:
Fresh manager sets the floor price the agent may not cross.
Agents:
Countdown Pricer, Demand Sensor
Reads:
Batch expiry, Hourly sales curve, Footfall
Writes:
Markdown ladder, Rescue basket, Donation handover

Never prices below the agreed floor; every step logged.

Open Fresh food and production

Fresh food and production

Production-to-demand planner

Live

Tells each kitchen what to bake, in what tray count, per two-hour slot.

Modelled value

AED 19m a year

Adoption

58%

Owner:
Production manager
Humans in the loop:
Head baker accepts or edits the plan each shift.
Agents:
Production Planner, Weather Signal
Reads:
Hourly sales, Weather, Local events, Ramadan calendar
Writes:
Tray plan, Ingredient call-off

Plans are proposals; the kitchen signs the final sheet.

Open Fresh food and production

Fresh food and production

Short-dated delivery rescue

Piloting

Offers short-dated fresh to nearby delivery baskets before it is lost.

Modelled value

AED 9m a year

Adoption

26%

Owner:
E-commerce and fresh
Humans in the loop:
Fresh manager releases each rescue window.
Agents:
Rescue Router, Proximity Offer Agent
Reads:
Batch expiry, Live delivery orders, Rider radius
Writes:
Rescue offer, Rider batch

Only lines inside food-safety windows.

Open Fresh food and production

Fresh food and production

Cold chain sentinel

Building

Watches chiller and freezer traces and calls a break before product is lost.

Modelled value

AED 11m a year

Adoption

9%

Owner:
Quality and safety
Humans in the loop:
Technician confirms and logs the fix.
Agents:
Cold Chain Sentinel
Reads:
Chiller probes, Door events, Maintenance history
Writes:
Break alert, Product hold

Product holds need a named quality sign-off.

Open Fresh food and production

Supply chain and logistics

Autonomous replenishment

Live

Orders to shelf-truth demand rather than book stock.

Modelled value

AED 54m a year

Adoption

69%

Owner:
Supply chain director
Humans in the loop:
Planners hold veto on any order above threshold.
Agents:
Replenishment Planner, Demand Sensor, Shelf Sentinel
Reads:
Shelf truth, DC stock, Lead times, Promotion calendar
Writes:
Purchase order, Store allocation

Orders above AED 250k route to a human planner.

Open Supply chain and logistics

Supply chain and logistics

Load and route builder

Piloting

Builds drops by store window, heat and traffic each evening.

Modelled value

AED 18m a year

Adoption

34%

Owner:
Logistics head
Humans in the loop:
Transport controller releases the plan.
Agents:
Route Builder, Traffic Signal
Reads:
Order lines, Vehicle capacity, Store windows, Traffic
Writes:
Load plan, Driver run sheet

Driver hours rules are hard limits, not preferences.

Open Supply chain and logistics

Supply chain and logistics

Inbound exception desk

Live

Chases suppliers in writing when an inbound slips.

Modelled value

AED 13m a year

Adoption

61%

Owner:
Vendor management
Humans in the loop:
Buyer approves any claim or penalty.
Agents:
Supplier Correspondent, Claims Agent
Reads:
ASN feed, Gate scans, Contract terms
Writes:
Written chase, Claim draft, Escalation

No claim is filed without a buyer's signature.

Open Supply chain and logistics

Supply chain and logistics

Import and customs copilot

Designed

Reads documents, pre-fills declarations, flags mismatches.

Modelled value

AED 9m a year

Adoption

0%

Owner:
Imports
Humans in the loop:
Customs broker signs every declaration.
Agents:
Document Reader, Compliance Checker
Reads:
Invoices, Packing lists, HS master
Writes:
Draft declaration, Mismatch list

Never files; drafts only.

Open Supply chain and logistics

E-commerce and delivery

Substitution negotiator

Live

Offers a substitute the shopper actually accepts, in their language.

Modelled value

AED 22m a year

Adoption

76%

Owner:
E-commerce director
Humans in the loop:
Picker confirms in the aisle.
Agents:
Substitution Agent, Aisle Concierge
Reads:
Live stock, Past acceptance, Price parity rules
Writes:
Substitute offer, Refund decision

Never upsells price without consent.

Open E-commerce and delivery

E-commerce and delivery

Slot and promise engine

Live

Promises only the slots the network can actually serve.

Modelled value

AED 16m a year

Adoption

67%

Owner:
Fulfilment
Humans in the loop:
Node manager can close a slot instantly.
Agents:
Slot Planner, Rider Broker
Reads:
Pick capacity, Rider supply, Traffic
Writes:
Open slots, Surge flag

Promise never exceeds observed capacity.

Open E-commerce and delivery

E-commerce and delivery

Basket build assistant

Piloting

Rebuilds a shopper's usual basket in one tap with fresh timing.

Modelled value

AED 19m a year

Adoption

29%

Owner:
Digital merchandising
Humans in the loop:
Merchandiser approves every promoted swap.
Agents:
Basket Agent, Loyalty Brain
Reads:
Order history, Loyalty profile, Stock
Writes:
Prefilled basket, Reminder

Suppression and consent honoured before any nudge.

Open E-commerce and delivery

E-commerce and delivery

Order care desk

Live

Handles where-is-my-order and refunds in writing, end to end.

Modelled value

AED 8m a year

Adoption

72%

Owner:
Customer care
Humans in the loop:
Care lead owns anything above AED 300.
Agents:
Care Agent, Refund Agent
Reads:
Order state, Rider position, Refund policy
Writes:
Written answer, Refund draft

Refunds above threshold need a human.

Open E-commerce and delivery

Merchandising, pricing and margin

Price position watcher

Live

Compares shelf price to market daily and proposes a move with a reason.

Modelled value

AED 62m a year

Adoption

66%

Owner:
Group merchandising
Humans in the loop:
Category head approves each price move.
Agents:
Price Watcher, Competitor Reader
Reads:
Own prices, Market scrapes, Elasticity history
Writes:
Price move proposal, Hold decision

No automatic price change; proposals only.

Open Merchandising, pricing and margin

Merchandising, pricing and margin

Margin leak detector

Live

Finds where margin leaves quietly: wrong cost, missed claim, bad promo.

Modelled value

AED 38m a year

Adoption

54%

Owner:
Commercial finance
Humans in the loop:
Finance signs off each recovery.
Agents:
Leak Detector, Claims Agent
Reads:
Cost master, Invoices, Promotion funding, Rebate terms
Writes:
Leak case, Recovery claim

Every case shows the invoice line it came from.

Open Merchandising, pricing and margin

Merchandising, pricing and margin

Supplier terms negotiator

Piloting

Prepares a negotiation pack and drafts the written ask.

Modelled value

AED 25m a year

Adoption

22%

Owner:
Buying director
Humans in the loop:
Buyer runs the conversation; agent drafts only.
Agents:
Supplier Correspondent, Terms Analyst
Reads:
Trading terms, Volume history, Category benchmarks
Writes:
Negotiation pack, Draft letter

Never commits Lulu to terms.

Open Merchandising, pricing and margin

Merchandising, pricing and margin

Range and listing candidate engine

Building

Proposes listings and delistings from demand the network refused.

Modelled value

AED 16m a year

Adoption

14%

Owner:
Category head
Humans in the loop:
Range committee decides.
Agents:
Range Agent, Unserved Demand Agent
Reads:
Unserved demand log, Search misses, Space plan
Writes:
Listing candidates, Delist candidates

Committee decision recorded against each candidate.

Open Merchandising, pricing and margin

Customer, loyalty and marketing

Dormancy recovery agent

Live

Writes to lapsing members with a reason they recognise.

Modelled value

AED 32m a year

Adoption

74%

Owner:
Loyalty director
Humans in the loop:
CRM lead approves each wave.
Agents:
Loyalty Brain, Outbound Voice Agent, Copy Agent
Reads:
Member history, Category lapse, Consent register
Writes:
Written offer, Call task, Suppression update

Consent and suppression checked before every send.

Open Customer, loyalty and marketing

Customer, loyalty and marketing

Live voice concierge

Live

Inbound and outbound voice that books, confirms and follows up.

Modelled value

AED 18m a year

Adoption

58%

Owner:
Customer experience
Humans in the loop:
Store team picks up any refusal within 60 seconds.
Agents:
Inbound Voice Agent, Outbound Voice Agent
Reads:
Member record, Store diary, Order state
Writes:
Call summary, Task, Written follow-up

Agent states it is an assistant and hands over on request.

Open Customer, loyalty and marketing

Customer, loyalty and marketing

Public listening desk

Live

Reads public posts and reviews per store and routes what matters.

Modelled value

AED 9m a year

Adoption

62%

Owner:
Brand and comms
Humans in the loop:
Store manager answers anything named.
Agents:
Listening Agent, Reputation Router
Reads:
Reviews, Public social, Store mentions
Writes:
Store task, Theme report

No public reply without a named human.

Open Customer, loyalty and marketing

Customer, loyalty and marketing

Campaign studio

Piloting

Builds local creative, reels and copy from live category signals.

Modelled value

AED 15m a year

Adoption

36%

Owner:
Marketing director
Humans in the loop:
Brand approves every asset before it runs.
Agents:
Copy Agent, Reel Agent, Spend Agent
Reads:
Category signals, Brand kit, Spend plan
Writes:
Creative set, Media plan, Attribution read

Brand sign-off is mandatory and logged.

Open Customer, loyalty and marketing

People and workforce

Screening and interview agent

Live

Screens and books in five languages, never scores a person alone.

Modelled value

AED 13m a year

Adoption

61%

Owner:
Talent acquisition
Humans in the loop:
Hiring manager makes every decision.
Agents:
Screening Agent, Scheduler
Reads:
Applications, Role profiles, Roster gaps
Writes:
Shortlist, Interview slots

No automated rejection; agent explains and a human decides.

Open People and workforce

People and workforce

Roster broker

Piloting

Fills tomorrow's gaps with willing, qualified colleagues.

Modelled value

AED 16m a year

Adoption

33%

Owner:
Store operations
Humans in the loop:
Supervisor confirms every shift change.
Agents:
Roster Broker, Queue Forecaster
Reads:
Roster, Availability, Skills, Labour rules
Writes:
Shift offer, Coverage plan

Labour law limits are hard constraints.

Open People and workforce

People and workforce

Attrition early warning

Live

Names the teams at risk and what has already been tried.

Modelled value

AED 21m a year

Adoption

49%

Owner:
HR director
Humans in the loop:
Line manager owns the conversation.
Agents:
Retention Sensor
Reads:
Tenure, Shift patterns, Pay events, Exit reasons
Writes:
Risk list, Suggested conversation

No individual score is shown without HR context.

Open People and workforce

People and workforce

First ninety days copilot

Building

Walks a new joiner through training, kit and check-ins.

Modelled value

AED 7m a year

Adoption

11%

Owner:
Learning
Humans in the loop:
Buddy and supervisor own the check-ins.
Agents:
Onboarding Agent
Reads:
Joiner record, Training plan
Writes:
Day plan, Check-in prompt

Training completion still recorded by a human.

Open People and workforce

Property, energy and facilities

Energy load agent

Piloting

Shifts chillers, lighting and HVAC against tariff and footfall.

Modelled value

AED 26m a year

Adoption

27%

Owner:
Facilities director
Humans in the loop:
Site engineer approves any setpoint change.
Agents:
Energy Agent, Weather Signal
Reads:
Meter reads, Tariff, Weather, Footfall
Writes:
Setpoint proposal, Load shift plan

Comfort and food-safety limits are hard bounds.

Open Property, energy and facilities

Property, energy and facilities

Predictive maintenance

Building

Calls a failure before it closes an aisle or a till.

Modelled value

AED 15m a year

Adoption

12%

Owner:
Engineering
Humans in the loop:
Technician confirms the fault and the fix.
Agents:
Asset Sensor, Cold Chain Sentinel
Reads:
Asset telemetry, Work orders, Parts stock
Writes:
Work order, Parts call-off

Work orders always name a responsible engineer.

Open Property, energy and facilities

Property, energy and facilities

Lease and tenant copilot

Designed

Reads leases, models renewals, drafts tenant correspondence.

Modelled value

AED 12m a year

Adoption

0%

Owner:
Property
Humans in the loop:
Leasing manager negotiates and signs.
Agents:
Document Reader, Terms Analyst
Reads:
Leases, Tenant sales, Footfall
Writes:
Renewal model, Draft letter

Drafts only; never binds Lulu.

Open Property, energy and facilities

Finance, risk and shared services

Close copilot

Piloting

Reconciles, explains variances and drafts the commentary.

Modelled value

AED 18m a year

Adoption

30%

Owner:
Group finance
Humans in the loop:
Controller signs the close.
Agents:
Reconciliation Agent, Narrative Agent
Reads:
Ledgers, Bank feeds, Prior commentary
Writes:
Variance pack, Draft commentary

No journal is posted by an agent.

Open Finance, risk and shared services

Finance, risk and shared services

Shrink and fraud sensor

Live

Joins till, camera and stock evidence into named cases.

Modelled value

AED 34m a year

Adoption

52%

Owner:
Loss prevention
Humans in the loop:
LP manager owns every case and interview.
Agents:
Shrink Sensor, Shelf Sentinel
Reads:
Till exceptions, Camera events, Stock counts
Writes:
Case file, Coaching task

No accusation; evidence goes to a named human only.

Open Finance, risk and shared services

Finance, risk and shared services

Touchless invoice desk

Live

Matches, queries and routes invoices without a keyer.

Modelled value

AED 15m a year

Adoption

71%

Owner:
Shared services
Humans in the loop:
AP lead resolves every mismatch.
Agents:
Document Reader, Reconciliation Agent
Reads:
Invoices, POs, Receipts
Writes:
Matched invoice, Query letter

Payment release stays with finance.

Open Finance, risk and shared services

Finance, risk and shared services

Cash and FX assistant

Designed

Forecasts cash by country and flags exposure early.

Modelled value

AED 10m a year

Adoption

0%

Owner:
Treasury
Humans in the loop:
Treasurer decides every position.
Agents:
Cash Forecaster
Reads:
Bank feeds, Sales forecast, Payment runs
Writes:
Cash view, Exposure flag

Advisory only; no transactions.

Open Finance, risk and shared services

Private label and sourcing

Own brand gap finder

Piloting

Finds categories where an own brand would win, with the evidence.

Modelled value

AED 29m a year

Adoption

24%

Owner:
Private label director
Humans in the loop:
Brand committee decides each launch.
Agents:
Range Agent, Competitor Reader
Reads:
Category sales, Price gaps, Market data
Writes:
Launch candidate, Price architecture

Committee decision recorded per candidate.

Open Private label and sourcing

Private label and sourcing

Supplier quality reader

Live

Reads test reports and audits, flags what fails before it ships.

Modelled value

AED 14m a year

Adoption

57%

Owner:
Quality assurance
Humans in the loop:
QA manager blocks or releases.
Agents:
Document Reader, Compliance Checker
Reads:
Lab reports, Audit files, Spec master
Writes:
Pass or block, Corrective action

Only a QA manager may release a blocked batch.

Open Private label and sourcing

Private label and sourcing

Sourcing scout

Building

Shortlists suppliers against spec, cost and compliance.

Modelled value

AED 18m a year

Adoption

16%

Owner:
Global sourcing
Humans in the loop:
Sourcing lead negotiates.
Agents:
Sourcing Agent, Terms Analyst
Reads:
Supplier database, Specs, Freight rates
Writes:
Shortlist, Landed cost model

No supplier is approved without an audit on file.

Open Private label and sourcing

Private label and sourcing

Pack and label compliance

Designed

Checks artwork against each country's labelling rules.

Modelled value

AED 7m a year

Adoption

0%

Owner:
Regulatory
Humans in the loop:
Regulatory affairs signs artwork.
Agents:
Compliance Checker, Copy Agent
Reads:
Artwork, Country rules, Ingredient master
Writes:
Compliance report, Correction list

Artwork release stays with regulatory.

Open Private label and sourcing