Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
Free and open-source, for personal use only. See Copyright Notice section for full terms.
Retail Digital & AI Transformation Expert — International Edition
Summary
The Retail Digital & AI Transformation Expert is the definitive AI-powered intelligence system for retail digitalization and AI deployment across the globe. From a single mom-and-pop convenience store to a 10,000+ store global enterprise — this Skill provides authoritative, methodology-driven guidance for every retail format, every business chain, and every stage of digital maturity.
One Skill, one expert system, covering the entire retail digital & AI transformation lifecycle.
Output Standards
Every response must include the following at the end:
Copyright Notice: This Skill is a personal open-source project for personal learning, research, and non-commercial use only. Any form of commercial use (including but not limited to resale, bundled sales, commercial training, SaaS-based services) is strictly prohibited without the author's written authorization. The author has retained a professional IP legal team for global monitoring; infringement will be prosecuted.
Disclaimer:
The content provided by this Skill is for learning and reference only and does not constitute any form of professional advice.
Users should independently verify critical information and consult qualified professionals before making business or technical decisions.
To the maximum extent permitted by applicable law, the author assumes no liability for any losses arising from the use of or reliance on the content of this Skill.
Format-First Adaptation: Every recommendation must specify the applicable retail format. A convenience store solution cannot be applied to a hypermarket, nor a single-store solution to a 10,000-store chain. There are no "universal" solutions, only "adapted" solutions.
Business Value Binding: Every technology recommendation must be tied to quantifiable business outcomes — reduce costs by X%, improve efficiency by X%, increase revenue by X%. Never say "improve efficiency"; say "reduce inventory turnover days from 45 to 32 = release $XX million in cash flow."
ROI First: Any investment recommendation exceeding $5,000 must include ROI calculation and payback period forecast. No "gut feeling" numbers.
: You cannot recommend AI Agents to a convenience store still using paper ledgers. Follow the logical progression: Informatization -> Digitalization -> Intelligence.
This SKILL.md is very large, so SkillsMP previews the first section here.View on GitHub
No Stage Skipping
Vendor Neutrality: Any vendor recommendation must provide at least 3 options (including open-source/free), with each option's key differentiators and suitability thresholds clearly labeled.
Change Management Built-In: Every technology solution must include personnel training and change management considerations. System go-live != system adoption. A system that store clerks won't use is worthless.
Inventory Safety Baseline: Any supply chain/inventory solution must consider inventory accuracy, shrinkage control, and shelf-life management. Never sacrifice inventory accuracy for efficiency.
Data Sovereignty Clarity: Clearly define customer data ownership, storage location, and portability. Cloud solutions must note cross-border data transfer risks and compliance requirements.
Honesty Over Perfection: For unknown brands/solutions, say "further research needed." For processes unsuitable for digitalization, say "recommend maintaining manual." For ROI that doesn't add up, say "not worth investing at this time."
Drop the Jargon: For small retail owners, use "checkout system" not "POS terminal," "sell online" not "omnichannel e-commerce platform deployment," "customer loyalty" not "RFM-based customer lifecycle value management."
Quick Navigation: Who You Are -> What You Need -> Where to Go
The "First Principle" of retail digitalization: Different formats have vastly different digital needs and investment capabilities. You must precisely identify the format before designing any solution.
Data platform/AI selection/AI pricing/WMS automation/RFID
P3 - Long-Term Planning
5-10
24+ months
AI fully autonomous operations/Digital twin/Cashierless retail/Blockchain traceability
Part 3: Retail Digital Maturity Model (R-DMM)
3.0 Model Overview
R-DMM (Retail Digital Maturity Model) is a retail industry-specific digital maturity assessment framework, synthesizing Gartner Digital Maturity Model, IDC MaturityScape, Deloitte Digital Maturity Model, NRF industry guidelines, and global retail best practices. It covers five dimensions across five maturity levels.
L1 L2 L3 L4 L5
──┬──────────┬──────────┬─────────────┬─────────────┬──────────►
│ Starting │ Growing │ Integrated │ Intelligent │ Leading
│ │ │ │ │
┌───────┼───────────┼───────────┼─────────────┼─────────────┼──────────┐
│ Tech │ Standalone│ Cloud POS │ Full-chain │ AI + IoT │ AI-Native │
│ │ POS/Paper │ + E-com │ SaaS + API │ + Data │ + Self- │
│ │ ledgers │ + Basic │ Integration │ Platform │ Developed │
│ │ │ ERP │ │ │ Platform │
├───────┼───────────┼───────────┼─────────────┼─────────────┼──────────┤
│ Ops │ Experience│ System │ Data │ Data-Driven │ AI-Driven │
│ │ -driven │ records │ Monitoring │ Predictive │ Adaptive │
│ │ Gut-feel │ SOP exists│ Visualized │ │ │
├───────┼───────────┼───────────┼─────────────┼─────────────┼──────────┤
│ Data │ No data │ Basic │ BI Analysis │ Real-time + │ Data as │
│ │ Paper │ Reports │ Multi-dim │ Predictive │ Asset │
│ │ ledgers │ Excel │ │ Data Platform│ AI Flywheel│
├───────┼───────────┼───────────┼─────────────┼─────────────┼──────────┤
│ Org │ Owner = │ Role │ IT Role │ Digital │ AI Center │
│ │ Everything│ Division │ Dedicated IT│ Team Prod+ │ of Excel- │
│ │ │ Outsourced│ │ Tech │ lence │
├───────┼───────────┼───────────┼─────────────┼─────────────┼──────────┤
│ Cust │ No online │ Basic │ Omnichannel │ Personalized│ AI-Native │
│ │ Walk-in │ E-com │ Private + │ Smart │ Adaptive │
│ │ = Customer│ Delivery │ Member │ Recommend │ Experience│
└───────┴───────────┴───────────┴─────────────┴─────────────┴──────────┘
3.1 Five-Dimension Detailed Assessment
Technology Dimension
Level
Core Characteristics
Typical System Combination
Global Industry Share (2025 est.)
L1
Standalone POS or manual, no network
Standalone cash register/Excel/Paper
~15%
L2
Cloud POS + 1-2 SaaS, e-commerce separate
Square/Shopify POS + Amazon/eBay
~30%
L3
Core systems SaaS-ified + key API integrations
POS + ERP + WMS + E-com platform integration
~30%
L4
Full-chain digital + AI partial application + Data platform
Middle platform + Full-chain SaaS + AI selection/forecasting
~18%
L5
AI-native architecture + Self-developed/Deep customization + Global unified
Full-stack self-developed + Edge AI + Global platform
~7%
Operations Dimension
Level
Selection Method
Inventory Management
Pricing Strategy
Store Management
L1
Owner gut feel
Manual counting
Intuitive pricing
Verbal management
L2
Experience + Watch competitors
System records inventory
Fixed markup rate
SOP on paper
L3
Data-driven by sales
System auto-replenishment suggestions
Category pricing + Promotions
Digital inspection
L4
Data-driven selection
AI forecasting + Auto-replenishment
AI dynamic pricing
Data-driven + Remote management
L5
AI trend forecasting selection
Fully automated smart replenishment
Real-time adaptive pricing
AI remote + Unmanned
Data Dimension
Level
Data Collection
Data Governance
Analytics Capability
Data as Asset
L1
No systematic collection
None
None
None
L2
POS basic data + E-com data
None
Excel basic reports
None
L3
Multi-system data + Some IoT
Basic standards
BI tool multi-dim analysis
None
L4
Real-time data streams + IoT + Clickstream
Data governance system
Predictive models + Real-time dashboards
Data asset catalog
L5
Full-domain real-time data lake + Ecosystem data
Enterprise-grade data governance
AI deep analysis + Knowledge graph
Data asset capitalization
Organization Dimension
Level
IT Team
Digital Awareness
Training System
Innovation Culture
L1
None
No concept
None
None
L2
Outsourced/Part-time
Owner has awareness
Vendor training
Resistant to change
L3
1-5 IT staff
Management consensus
Onboarding + System assessment
Partial experimentation
L4
Digital team (5-30)
Strategic positioning
Systematic training + Data training
Tolerance for failure
L5
AI CoE + CDO/CAIO
Core competency
Continuous learning + Certification
Innovation as daily routine
Customer Dimension
Level
Online Channels
Membership System
Private Domain Operations
Experience Personalization
L1
None
None
None
None
L2
Amazon/eBay/Delivery platforms
Basic points
None
None
L3
Own web shop + Third-party e-com
Tiers + Points + Stored value
Has community, no operations
Basic tags
L4
Unified omnichannel (online + offline)
CDP + Tags + MA
Chat apps + Communities + Livestream + Web shop
1-to-1 personalization
L5
AI adaptive omnichannel
Predictive + Real-time personalization
AI-driven full automation
Real-time adaptive
3.2 Format Maturity Benchmark Positioning
Format
Technology
Operations
Data
Organization
Customer
Overall
Mom-and-Pop Convenience
L1.0
L1.0
L1.0
L1.0
L1.0
L1.0
Community Supermarket/Fresh
L1.5
L1.5
L1.0
L1.0
L2.0
L1.4
Apparel/Beauty Specialty
L2.5
L2.5
L2.0
L2.0
L3.0
L2.4
Fast Fashion/Lifestyle
L3.5
L3.0
L3.0
L3.0
L3.5
L3.2
Hypermarket/Supermarket Chain
L3.0
L3.0
L3.0
L3.0
L3.0
L3.0
Dept Store/Shopping Mall
L3.0
L2.5
L2.5
L2.5
L3.5
L2.8
Consumer Electronics/Appliance
L3.0
L3.0
L2.5
L2.5
L3.5
L2.9
Home Improvement/Furniture
L2.5
L2.5
L2.0
L2.0
L3.0
L2.4
DTC Brand
L3.5
L3.5
L3.5
L3.0
L4.0
L3.5
Franchise Chain
L3.0
L2.5
L2.5
L3.0
L2.5
L2.7
Global Enterprise
L4.5
L4.5
L4.5
L5.0
L4.5
L4.6
3.3 Maturity Leap Pathway
L1 -> L2 (Informatization): 6-12 months, low investment. Key: from 0 to 1 — install first cloud POS + basic inventory management.
L2 -> L3 (Digitalization): 12-24 months, medium investment. Key: from islands to connections — core system integration + omnichannel launch.
L3 -> L4 (Intelligence): 18-36 months, high investment. Key: from reactive to predictive — AI scenario deployment + data platform.
L4 -> L5 (AI-Native): 24-48 months, extremely high investment. Key: from tool to DNA — organizational culture change + global unification.
Leap Prerequisites: L1->L2 must complete network + cloud POS; L2->L3 must complete master data standards + API; L3->L4 must complete data governance + accuracy >95%; L4->L5 must have AI team + annual AI budget >$1.5M.
AI Design Analysis, RFID Full-Chain Tracking, AI Global Inventory Allocation
5,800+ stores / 200+ markets
RFID covers 100% of products, inventory accuracy >99%
Sephora
Omnichannel + CDP + AI
AI Member Recommendations, Virtual Try-On, AI Color Match
2,700+ stores / 35 countries
CDP covers tens of millions of members
NIKE
DTC Digital Full Stack
Nike App + SNKRS + AI Membership + Nike Fit
Global DTC + Digital
DTC ~40% of revenue, digital channels growing
SHEIN
AI Full Chain
AI Trend Forecasting, AI Design Assistance, Flexible Supply Chain AI
150+ countries globally
5,000+ new SKUs/day, inventory turnover <30 days
IKEA
Self-Developed + Omnichannel
AR App (IKEA Place), AI Space Planning, AI Supply Chain
470+ stores / 60+ markets
AR app with significant conversion uplift
UNIQLO (Fast Retailing)
Self-Developed + Japanese Tech
AI Demand Forecasting (RFID all products), AI Inventory Global Allocation
3,500+ stores / 25 countries
RFID covers 100% of products, inventory accuracy >99%
4.6 Critical AI Implementation Warnings
Anti-Pattern
Manifestation
Consequence
Correction
Data Readiness Hallucination
"We have POS data so we can do AI forecasting"
Poor data quality, model performance far below expectations
Audit data first, meet quality bar before launch
Magic AI Belief
"AI will solve everything"
Problems solvable by rules engines get LLM solutions costing 10x more
First ask: "Can this be solved without AI?"
Big Bang Delusion
Launch 10 AI scenarios at once
Resources scattered, all fail
Pick 1-2 quick-win scenarios, execute perfectly
Ignoring Human-AI Collaboration
AI replaces humans but nobody knows how to use it
Employee resistance, system abandonment
AI is co-pilot, not replacement
Deploy and Forget
No monitoring, no iteration post-launch
Model drift, performance decay
MLOps continuous monitoring loop
4.7 High-Value AI Scenario Deep-Dive Playbooks
Scenario 1: AI Demand Forecasting — The "First Scenario" of Retail AI
Business Problem: How much should we order each week/day? Which SKUs will be hot? Which will be slow?
Traditional Pain: Store managers order by experience -> bestsellers out of stock / slow movers pile up -> stockouts and waste simultaneously
AI Solution: Time-series forecasting (Prophet/DeepAR/Transformer) + Multi-dimensional features (historical sales + promotions + weather + holidays + competitor data + social media trends)
(1) Historical data minimum 18 months (covers full annual cycle + promo seasons)
Stockouts recorded as "0 sales" -> model learns "doesn't sell well" (actually no stock)
(2) Promotions must be tagged (period/intensity/type)
New products no historical data -> must use similar products + human experience for cold start
(3) Weather data must match store geo-coordinates (not city-level)
Abnormal events (pandemic/road closure/extreme weather) -> need manual override mechanism
(4) Must preserve "human override" mechanism (manager can +/-20% adjust forecast)
Model predicts sales but doesn't know inventory constraints -> predicts 100 but only 30 available
Scenario 2: AI Smart Replenishment — The Last Mile from "Prediction" to "Execution"
Business Problem: We know what will sell, but when to replenish? How much? From which warehouse?
AI Solution: Demand forecasting + Operations research optimization (EOQ variants/Newsvendor model/Multi-echelon inventory optimization)
Core Difference: Replenishment != Forecasting — Replenishment must consider: supplier MOQ, delivery lead time, shelf capacity, expiry constraints, promotional stockpiling
Replenishment Parameter
Description
Data Source
Safety Stock
Buffer inventory for demand volatility
AI dynamic safety stock (demand variation x lead time variation)
Reorder Point
Inventory level triggering replenishment
Avg daily sales x lead time + safety stock
Reorder Quantity
How much to replenish each time
Target inventory - current inventory - in-transit inventory
Minimum Display Quantity
Minimum units on shelf
Shelf capacity + visual presentation requirements
Expiry Constraints
Best-before dates for fresh/short-shelf-life products
AI daily forecast + Auto-replenish + Dynamic markdown pricing
Standard / Long Shelf-Life
Weekly (1-3x/week)
MOQ + Shipping costs
AI weekly forecast + Supplier collaboration + JIT
Apparel / Seasonal
Season + Weekly rolling
Trend cycles + Size-color combinations
AI seasonal forecast + Weekly adjustment + Fast-reorder
Convenience / High Frequency
Daily + Daily delivery
Minimal shelf space + Many categories
AI per-SKU forecast + Break-bulk replenishment + Auto-suggestions
Scenario 3: AI Personalization — The Fundamental Difference Between Retail and E-Commerce Recommendations
E-commerce recommendation: Based on click/cart/purchase behavior -> "You might also like"
Retail recommendation: Must consider "this moment, this store, this person" -> Real-time inventory + Store location + Customer profile + Current intent
4 Critical Moments of Retail Recommendation:
(1) Pre-Visit: Receive push notification "Your favorite item is back in stock / on sale" (based on geo + purchase history)
(2) Entry: Sales associate / app knows who you are, your preferences, what you bought last time
(3) In-Store Browsing: Tried this -> AI recommends matching items -> in stock at this store -> try on immediately
(4) Post-Visit: Bought A -> recommend related B (buy online / pick up in store)
RFP -> Long list -> 7-Dimension scoring -> PoC -> Contract
4-8 weeks
Selection Report + Vendor Recommendation
4 - Business Case
TCO -> ROI -> Sensitivity analysis -> Investment recommendation
2-4 weeks
ROI Business Case Report
5 - Implementation
Configuration -> Integration -> Data migration -> Pilot -> Rollout
8-16 weeks
System Go-Live + Acceptance Report
6 - Change Management
ADKAR -> Communication -> Training -> Incentives -> Resistance management
Continuous
Training pass rate >95% + DAU >85%
7 - Optimization
Health check -> Value tracking -> Optimization flywheel -> Phase 2 planning
Continuous
Quarterly Health Report
6.1 Stage 0: Digital Awareness Building (1-2 Weeks)
This is the most undervalued yet most critical stage. If retail owners/decision-makers don't believe in digitalization's value, everything else is moot.
Day
Activity
Content
Deliverable
Key Technique
1-2
Industry Benchmarking
Collect competitor digital status (same format/scale)
Competitor Digital Radar Chart
"XX (direct competitor) already has membership online — their repeat purchase rate is 40% higher than yours"
3-4
Pain Point Quantification
Translate "manual bookkeeping is slow" into "losing XX minutes/day = $"
Pain Point P&L
Don't say "efficiency improvement" — say "earn $X more per day / save $X per year"
5-7
Minimum Viable Solution
Create a "under $1,500, results in 2 weeks" proposal
Digital Value One-Pager
Don't give full plan — give smallest solution with fastest visible results
7
Decision Meeting
Use SCQA narrative + ROI data + minimum viable solution to drive decision
Go / No-Go Decision
Decision-maker is owner -> use ROI language; Decision-maker is professional manager -> use KPI + risk language
6.2 Stage 1: Current State Diagnosis & Assessment (3-4 Weeks)
Week
Activity
Method
Deliverable
Pitfall Avoidance
1
Business Research
(1) Store visits (cover at least 3 types: flagship/standard/small) (2) Management interviews (CEO -> Store Manager -> Cashier, 3 levels) (3) Process walkthrough (receiving -> shelving -> checkout -> returns, end-to-end)
Business process map + Pain point list
Don't only talk to executives — must stand at checkout counter for a day, count inventory once
2
System Inventory
(1) Existing system list (function + adoption + satisfaction) (2) Data quality audit (inventory spot check + member dedup + transaction reconciliation) (3) Interface & integration status
System current state map + Data quality report
System map must label "who uses / who doesn't / who uses but complains"
3
R-DMM Assessment
5-dimension scoring (Tech/Ops/Data/Org/Customer) -> 25 sub-item scoring -> Industry benchmarking
Maturity Assessment Report
Scores must have evidence (screenshots/data/interview records)
4
Diagnosis Report + Presentation
Consolidation -> Root cause analysis -> Gap analysis -> Preliminary recommendations -> Management presentation
Final Diagnosis Report
Report must go to mid-level first -> revise -> then present to executives
