Hyundai Glovis Executive Enablement • 2 Hours

AI-POWERED PRODUCTIVITY &
SMART SUPPLY CHAIN

From Generative AI & Prompt Engineering to Intelligent Logistics, Decision Intelligence & AI Agents

Automotive SCM Decision Support AI Agent Swarms Practical | Hands-on

💡 Core Paradigm: "AI is not just a tool. It is a new way of working."

Masterclass Agenda

🗺️ Today’s Learning Roadmap

02 / 58

01

AI & GenAI

Core concepts & foundational shifts

02

AI Tools

Tool taxonomy & selection heuristics

03

Prompting

7-Part formula & 7 power tricks

04

Productivity

Emails, Meetings, Reports & Excel

05

Smart SCM

Demand, Shipping, Routes & Warehouse

06

Decisions

Risk scoring & Scenario modeling

07

AI Agents

Autonomous loops & Digital workforce

08

Responsible AI

Data hygiene & 5-step verification

Strategic Context

🏢 Why AI Matters to Hyundai Glovis

03 / 58

Integrated Glovis Operational Footprint

Finished Vehicle (PCC) CKD / KD Logistics Global Shipping Smart Warehousing Inland Transport Air Cargo Forwarding EV Battery Logistics Future Mobility

Every operational node generates massive transactional and sensor streams ripe for continuous intelligence.

The Smart SCM Formula

Glovis Data + AI Models + SCM Expertise

= SMART SCM

Unlocking real-time visibility, predictive ETAs, and zero-touch exception handling.

Industry Paradigm

📈 The AI Evolution in Enterprise

04 / 58

Dimension	Yesterday (Traditional)	Today (GenAI Copilots)	Tomorrow (AI Agents)
Information Access	Keyword search & portal lookups	Conversational summaries & answers	Continuous background intelligence
Data Analysis	Manual Excel parsing & VLOOKUPs	AI-assisted formula & trend mining	Automated anomaly alerts & root cause
Operational Action	Manual ticket creation & emailing	Copilot drafts response for review	Autonomous execution within guardrails
Role of AI	Static Software Tool	Interactive Teammate / Copilot	Scalable Digital Workforce

Core Fundamentals

🤖 What is Artificial Intelligence?

05 / 58

Core Definition

AI enables computing systems to perform cognitive tasks that historically required human intelligence.

Read & SynthesizeUnderstand Context Analyse TrendsPredict Volumes Generate ArtifactsRecommend Trade-offs

⚠️ Non-Negotiable Operational Reality

AI outputs can be factually wrong.

AI is a probabilistic reasoning engine, not an infallible database. Critical operational decisions always require a human validation layer.

GenAI Mechanics

✨ What is Generative AI?

06 / 58

Multimodal Creative & Analytical Capabilities

Traditional AI categorized numbers; Generative AI synthesizes new structured artifacts:

Executive ReportsEmail Drafts JSON DataSOP Guides Python ScriptsSlide Outlines Disruption BriefsTranslations

Primary Engines: OpenAI GPT-4o, Google Gemini Pro, Anthropic Claude 3.5, MS Copilot.

The Iterative Generation Loop

1

Ask

Prompt

2

Gen

Draft

3

Review

Audit

4

Tune

Finalize

Enterprise Ecosystem

🗺️ Enterprise AI Tool Map

07 / 58

Operational Need	AI Tool Category	Industry Examples	Primary SCM Use Case
General Reasoning	AI Assistant	ChatGPT, Gemini, Claude	Brainstorming mitigation, draft emails, policy review
Microsoft Integration	Enterprise Copilot	Microsoft 365 Copilot	Excel variance modeling, Teams meeting action extraction
Real-time Research	Grounded Search AI	Perplexity, Gemini 1.5	Tracking port strikes, tariff changes, canal congestion
Long PDF Binders	Source-Grounded AI	NotebookLM, Acrobat AI	Vendor SLAs, customs documentation, maritime regulations
Process Automation	Agentic Workflows	Power Automate, Copilot Studio	Automated freight audit, demurrage exception triage

Comparative Analysis

⚖️ Model Matrix: ChatGPT vs Gemini vs Claude vs Copilot

08 / 58

Capability Area	ChatGPT Plus / Team	Google Gemini Advanced	Anthropic Claude 3.5	M365 Copilot
Data Analysis & Code	⭐⭐⭐⭐⭐ (Code Interpreter)	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐
Long Document Context	⭐⭐⭐⭐ (128k)	⭐⭐⭐⭐⭐ (2M Tokens)	⭐⭐⭐⭐⭐ (200k Artifacts)	⭐⭐⭐⭐
Complex Reasoning	⭐⭐⭐⭐⭐ (o1 / GPT-4o)	⭐⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐
Live Web Search	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐
Office 365 Workflow	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐ (Native)

Decision Framework

⏱️ The 5-Second Tool Selection Rule

09 / 58

Draft / Brainstorm / Code

ChatGPT / Claude

Best for unstructured strategic prompts, code generation, and rewriting.

Inside Office Apps

M365 Copilot

Best for working directly inside Excel spreadsheets, Word memos, and Outlook.

Live Web Intelligence

Perplexity / Gemini

Best for real-time geopolitical updates, port congestion, and tariff verification.

Multi-Document Binders

NotebookLM

Best for strict zero-hallucination Q&A grounded inside uploaded PDF manuals.

Instruction Precision

🎯 Prompt Engineering: Quality In = Quality Out

10 / 58

❌ Weak Prompt

"Analyse logistics."

Outcome: Generic textbook generalities with zero actionable business value.

⚠️ Better Prompt

"Analyse the attached logistics data and identify the top 5 delivery issues."

Outcome: Relevant bullet list, but lacks quantification and structural rigor.

✅ Production-Ready Prompt

"Act as a senior SCM analyst. Identify top 5 delivery bottlenecks, quantify variance, list root causes, and provide a decision table."

Outcome: Executive-ready decision artifact formatted for management review.

Prompt Architecture

📐 The 7-Part Prompt Formula

11 / 58

R

Role

Who is AI acting as?

C

Context

Business background

T

Task

Exact required action

D

Data

Input text / metrics

C

Constraint

Boundaries & limits

O

Output

Table / Memo / JSON

V

Verify

Flag uncertainties

Ready-to-Use Artifact

📋 Master Prompt Template

12 / 58

ENTERPRISE TEMPLATE // SCM OPERATIONS

Act as a [ROLE: Senior Automotive Supply Chain Specialist]. Context: [BUSINESS CONTEXT: Reviewing Q3 finished vehicle dispatch delays from Ulsan to Chennai]. Task: [TASK: Identify top 3 variance drivers, assess demurrage impact, and propose alternative routes]. Data: [ATTACHED DATASET / TRANSLOAD LOGS]. Constraints: Do not assume unstated transit times. Adhere to Glovis standard SLA buffer rules. Output: Provide a structured Markdown table: | Route Lane | Variance (Days) | Cost Risk | Root Cause | Proposed Action |. Verification: Clearly label unverified assumptions and data gaps in a separate notes section.

Prompt Trick 01

💡 Technique #1: Ask AI to Ask Questions First

13 / 58

Traditional Flawed Approach

"Create a warehouse slotting strategy for our automotive KD distribution center."

AI will hallucinate layout dimensions, velocity curves, and SKU counts.

REVERSE-INTERVIEW PROMPT

Act as a senior warehouse operations engineer. I want to optimize slotting for our KD facility. Instruction: Before proposing any layout, ask me the 7 most critical questions you need regarding throughput velocity, SKU profiles, forklift constraints, and dispatch peaks. Do not generate the strategy until I answer.

Prompt Trick 02

🎭 Technique #2: Assign Precise Professional Roles

14 / 58

ROLE-PRIMED PROMPT

Act as a Senior Procurement and Supply Chain Risk Director with 15+ years evaluating Tier-1 automotive parts suppliers in East Asia. Analyse this vendor financial audit and delivery scorecard. Focus strictly on sole-source exposure, lead-time volatility, and working capital solvency.

Why Role Priming Changes Output:

Filters out generic fluff: Uses precision industry terms (OTIF, Safety Stock buffers, DPO).
Prioritizes risk factors: Focuses on downside continuity rather than marketing metrics.
Adopts executive brevity: Formats results for high-level steering committees.

Prompt Trick 03

📊 Technique #3: Enforce Structured Tabular Outputs

15 / 58

SCHEMA PROMPT

Output Schema Preview

Node	OEM Impact	Risk	Action
Port Gate	Line stoppage	HIGH	Expedite customs
Transload	+24h Dwell	MED	Add 2nd shift

Prompt Trick 04

⚔️ Technique #4: Ask AI to Challenge You

16 / 58

DEVIL'S ADVOCATE PROMPT

Act as a Devil's Advocate and Stress-Testing Auditor. Here is our proposed logistics consolidation plan: [INSERT PLAN / ROUTE SHIFT]. Your Task: Ruthlessly challenge this proposal. 1. Identify hidden operational dependencies and failure points. 2. Highlight uncalculated cost blowouts (demurrage, chassis dwell, port tariffs). 3. Give me the single strongest argument for why this plan might fail in Q4.

Prompt Trick 05

🔍 Technique #5: Force AI to Show Uncertainty

17 / 58

UNCERTAINTY PROMPT

Analyse this new trade tariff proposal and its impact on our automotive exports. Structure your analysis into 4 distinct categories: 1. Verified Facts: (Supported directly by official text) 2. Operational Assumptions: (Reasonable extrapolations) 3. Calculated Estimates: (Estimated cost impacts) 4. Critical Unknowns: (Data missing; requires confirmation)

Why This Prevents Disasters:

LLMs naturally speak with 100% confidence even when guessing. Forcing uncertainty categorization guarantees that executives never confuse an AI extrapolation with verified tariff law.

Prompt Trick 06

🔀 Technique #6: Request Multi-Option Trade-Offs

18 / 58

Don't Ask:

"What should we do about the port congestion?"

AI will prescribe a single generic path without weighing financial trade-offs.

TRADE-OFF MATRIX PROMPT

Provide 3 viable strategic alternatives for handling the 5-day port congestion at destination. For each option, evaluate: - Financial Cost Impact | Lead Time Delta - Implementation Friction | Operational Failure Risk - Exact Trigger Condition (When should this option be chosen?)

Prompt Trick 07

👓 Technique #7: Use AI as a Second Reader

19 / 58

AUDIT PROMPT

Act as a meticulous contract auditor. Review the attached Carrier Service Level Agreement (SLA). Highlight: 1. Ambiguous delivery timeline definitions. 2. Contradictions between demurrage clauses and detention waivers. 3. Uncapped liabilities or missing force majeure parameters. 4. Top 5 clarification questions we must ask the carrier before signing.

High-Value Application Areas:

Maritime Freight Contracts: Spotting hidden surcharge clauses.
Warehouse SOPs: Finding missing exception escalation paths.
Board Briefings: Auditing numerical consistency across tables.

Personal Productivity

✉️ AI for Email Communication

20 / 58

EMAIL PROMPT

Rewrite the following email draft to an overseas OEM customer regarding a 2-day shipment delay. Tone: Highly professional, reassuring, and solution-focused. Constraints: Keep under 120 words. State the exact cause (customs system migration), the new delivery milestone, and the 2 active mitigation steps we have taken. Provide a clear, urgent subject line.

Everyday Email Workflows:

Summarize 20-thread chains

Extract action owners

Tone adjustment (Executive)

Multilingual Korean/English

Personal Productivity

🎙️ AI for Meeting Intelligence

21 / 58

TRANSCRIPT TO MATRIX

Convert this 45-minute operations meeting transcript into a structured briefing: 1. Key Decisions Approved (3 bullets max) 2. Action Matrix: | Action Item | Owner | Deadline | Dependency | 3. Unresolved Disagreements / Open Questions 4. Next Follow-Up Milestone

Generated Action Matrix

Action	Owner	Deadline
Re-route Incheon containers	Logistics	Fri 17:00
Audit demurrage waiver caps	Procurement	Mon 12:00

Executive Reporting

📑 AI for Management Reports

22 / 58

REPORT SYNTHESIS PROMPT

Act as a chief of staff for supply chain operations. Analyse this 30-page monthly logistics KPI report. Synthesize into a 1-page Executive Briefing: - Executive Summary: 3 bullet highlights on overall network health. - Top 3 Variance Drivers: Which lanes deviated most from budget and why? - Emerging Risk Radar: 3 operational trends that could harm Q4 margin. - 5 Scrutiny Questions: What specific questions should the VP ask lane managers?

Analytical Productivity

📊 AI for Excel & Spreadsheet Analytics

23 / 58

Complex Formulas

Generate nested XLOOKUP, LAMBDA, and INDEX/MATCH formulas instantly by describing the goal in plain Korean/English.

Data Cleansing

Standardize messy vendor names, inconsistent date formats, and broken container tracking codes automatically.

Variance Detection

Upload raw CSV files to ChatGPT/Copilot to run instant statistical anomaly tests and isolate cost outliers.

Python Automation

Ask AI to write complete Python scripts to consolidate 50 regional branch spreadsheets into one master sheet.

SCM Domain Map

🚚 Logistics AI Use-Case Ecosystem

24 / 58

SCM Operational Area	AI Capability	SCM Operational Area	AI Capability
Demand Planning	Seasonal multi-variable forecasting	Inland Distribution	Last-mile route & load balancing
Procurement / Sourcing	Supplier solvency & disruption scoring	Fleet Management	Predictive engine telemetry maintenance
Warehouse & DC	Dynamic velocity slotting & stock-outs	Customer Service	Real-time chatbot milestone tracking
Maritime / Shipping	Vessel turnaround & congestion prediction	Executive SCM	Automated cross-border board briefings
Finished Vehicle (PCC)	Yard space allocation & dwell optimization	Risk Governance	Geopolitical & trade restriction radar

Smart SCM Use Case

📈 AI for Automotive Demand Forecasting

25 / 58

DEMAND FORECAST PROMPT

Act as a senior automotive demand planner. Analyse 24 months of CKD module dispatch data alongside OEM production schedules. Generate: 1. A 3-month forward projection accounting for holiday plant shutdowns. 2. Sensitivity analysis for a ±10% variation in EV battery pack availability. 3. List of SKUs with highest stock-out risk.

Forecasting Variables Combined:

Historical dispatch runs OEM assembly schedules Seasonal factory shutdowns Port dwell variance

Smart SCM Use Case

🛡️ AI for Supplier Risk Intelligence

26 / 58

VENDOR AUDIT PROMPT

Act as a supplier risk analyst. Score this key automotive Tier-1 supplier on a 1-100 scale across: - Delivery OTIF Reliability | Quality Defect PPM - Working Capital Health | Geopolitical / Single-Site Exposure Requirement: Provide explicit evidence for each score and flag missing data.

Simulated Supplier Risk Card

Dimension	Score	Risk	Key Evidence
Delivery OTIF	62/100	HIGH	3 late dispatches in Q2
Solvency	88/100	LOW	Strong liquidity ratio
Geo Exposure	45/100	HIGH	100% single plant in coastal zone

Smart SCM Use Case

🚛 AI for Transport & Route Optimization

27 / 58

TRANSPORT PROMPT

Act as a logistics optimization specialist. Analyse the attached shipment dispatch logs for Lane: Busan ➔ Chennai. Identify: Recurring port dwell bottlenecks, route variances, and priority JIT exceptions. Output: Provide 3 specific re-routing options with trade-offs in transit time vs. freight cost.

🎯 Simulated AI Decision Artifact

Option	Route	Lead Time	Cost Delta
Primary	Busan ➔ Chennai Direct	14 Days	Baseline
Alt A	Busan ➔ Singapore ➔ Air	4 Days	+38% (High)
Alt B	Busan ➔ Colombo Feeder	11 Days	+6% (Low)

✔ Recommendation: Trigger Alt B for non-critical CKD; reserve Alt A for line-stop risks.

Smart SCM Use Case

🏬 AI for Warehouse & Inventory Optimization

28 / 58

WAREHOUSE PROMPT

Analyse this 10,000-SKU warehouse inventory log. 1. Categorize inventory into Fast (A), Medium (B), and Slow/Dead (C) velocity. 2. Recommend dynamic slotting reallocations to minimize forklift travel distance by 15%. 3. Identify obsolete parts incurring excess storage holding costs.

Core DC Impact Metrics:

Forklift travel distance -18% Picking error reduction Dead stock holding cost -12% Dock-to-stock cycle time

Smart SCM Use Case

🚢 AI for Maritime & Ocean Logistics

29 / 58

MARITIME ETA PROMPT

Act as a maritime fleet operations coordinator. Evaluate current vessel schedules across our Pure Car Carrier (PCC) network. Flag: High-risk ETA slips due to Singapore bunker wait times and Suez diversion. Recommend vessel speed adjustments to balance fuel burn vs. demurrage penalties.

Maritime Factors Synthesized:

Port Congestion Dwell: Real-time waiting queues at destination berths.
Vessel Speed Optimization: Fuel surcharge vs. delay penalty curve.
Tide & Canal Windows: Lock scheduling & transit availability.

High-Impact SCM Pattern

🎯 AI for Exception Triage & Filtering

30 / 58

1,000

Active Shipments

Continuous real-time GPS & carrier feed intake

17

Critical Exceptions

High delay risk threatening OEM line continuity

983

Automated Pass

Normal operating tolerance; zero manual review needed

Operational Paradigm Shift:

Traditional teams manually inspect hundreds of spreadsheet rows. AI-enabled logistics teams focus 100% of human problem-solving on the 17 critical exceptions.

Disruption Intelligence

🌐 AI for Geopolitical & Trade Disruption

31 / 58

GEOPOLITICAL RADAR

Analyse the recent Red Sea shipping diversions and new carbon border taxes (CBAM). Produce: 1. Impact assessment on Hyundai Glovis Asia-Europe PCC transit times. 2. Fuel surcharge cost escalation estimate per vehicle unit. 3. 3 operational hedging recommendations for contract renewals.

Chokepoint Risks

Suez, Panama Canal, Malacca Strait disruptions monitored 24/7.

Trade & Tariffs

Instant breakdown of complex customs tariff changes.

Strategic Resilience

🔀 Multi-Variable Scenario Planning

32 / 58

Scenario 1

Normal Ops Baseline

Scenario 2

Destination Port Strike

Scenario 3

Bunker Fuel +30%

Scenario 4

Tier-1 Supplier Stoppage

Scenario 5

Customs System Outage

SCENARIO MODELING PROMPT

Compare these 5 operational scenarios across: Total Landed Cost | Lead Time Delta | Assembly Line Stop Risk | Recommended Pre-emptive Mitigation. Present in an executive comparison matrix.

Decision Framework

⚖️ AI for Decision Support & Trade-Offs

33 / 58

Strategic Option	Operational Benefit	Financial Cost	Execution Risk	Difficulty	Recommendation
Option A: Emergency Air Charter	Guaranteed 48h JIT arrival	Extreme (+$180k)	Very Low	Low	Trigger ONLY if line halts
Option B: Feeder Port Re-route	Saves 4 days vs. direct congestion	Moderate (+8%)	Medium	Medium	Recommended Default
Option C: Factory Buffer Hold	Zero extra freight cost	Holding cost	High (Stock-out)	High	Reject (High OEM risk)

Human Decision: Leadership selects Option B with pre-authorization for Option A if transit exceeds 72h.

Autonomous Systems

🤖 What is an AI Agent?

34 / 58

Traditional Generative AI

Ask ➔ Answer

Single turn interaction. Requires a human to manually provide data, read output, copy to another app, and trigger each action.

Autonomous AI Agent

Goal ➔ Observe ➔ Reason ➔ Act ➔ Report

Given a high-level goal, the agent accesses APIs, verifies telemetry, executes sub-tasks, and escalates exceptions.

1

Observe

Polls ERP & AIS Feeds

2

Detect

Spots +3 Day Port Dwell

3

Reason

Assesses JIT Factory Impact

4

Propose

Prepares Re-routing Plan

5

Notify

Alerts Operations Lead

Agentic Deployment

🏢 Enterprise AI Agent Use Cases

35 / 58

Business Function	Autonomous Agent Name	Core Trigger & Autonomous Action	Human Gate
Ocean Operations	Vessel Delay Triage Agent	Detects berth delay > 48h ➔ drafts feeder alternative	Manager approves booking
Procurement	Freight Invoice Audit Agent	Matches carrier invoice vs contract rate ➔ flags discrepancies	Approves payout override
Warehousing	Dynamic Re-order Agent	Detects velocity spike ➔ calculates safety stock delta	Purchasing signs PO
Customer Service	Milestone Notification Agent	Pulls GPS transload events ➔ writes personalized update	Auto-sent within tolerances

Organizational Maturity

🚀 Evolution to a Digital SCM Workforce

36 / 58

1

Level 1: Assistant

Ad-hoc Q&A and translation in ChatGPT / Claude.

2

Level 2: Copilot

Embedded inside Excel, Teams, and Word workflows.

3

Level 3: Automation

Deterministic multi-step flows in Power Automate.

4

Level 4: AI Agent

Continuous reasoning, monitoring & self-correction.

Risk & Ethics

🛡️ Responsible AI: Trust But Verify

37 / 58

Common Enterprise AI Vulnerabilities:

Hallucinated tariffs & citations Confidential data leakage Over-reliance on unverified math Algorithmic bias in vendor scoring

The Golden Operating Standard:

Trust ➔ But Verify

Use AI to draft, structure, compute, and model at 10x speed. Never deploy an output externally without professional validation.

Data Security

🚫 What Should NOT Go Into Public AI?

38 / 58

❌ Strictly Prohibited on Public Cloud AI:

Proprietary OEM customer pricing & negotiated contract tiers
Customer PII (Personal Identifiable Information) & phone records
System access credentials, database passwords & API keys
Unannounced quarterly business strategy & financial filings

Pre-Upload Authorization Test:

"Is this specific data category authorized to be processed by this AI system under current Hyundai Glovis IT security policy?"

Risk Governance

🔁 The 5-Step AI Verification Loop

39 / 58

1

Check

Are numbers & rates factually accurate?

2

Verify

Can ERP / carrier contract confirm it?

3

Challenge

What assumptions could fail in reality?

4

Review

Does it align with Glovis business logic?

5

Decide

Human manager executes final approval

Non-Negotiable Enterprise Principle:

AI generates options, calculates probabilities, and structures data. Accountability for financial spend, supplier commitments, and shipment dispatch remains 100% with the human professional.

Quality Assurance

✅ Prompt Quality Pre-Flight Checklist

40 / 58

Checklist (Part 1)

🎯 Clear Task: Did I state an exact operational verb?
📌 Context: Did I provide necessary background details?
📊 Data: Did I attach the clean dataset or transload log?
🚫 Constraints: Did I state what NOT to assume?

Checklist (Part 2)

📋 Output Format: Did I specify table headers / bullet limits?
🔍 Uncertainty: Did I instruct AI to separate facts from guesses?
👤 Human Gate: Did I retain human review before dispatch?

Prompt Vocabulary

⚡ 10 Power Commands for SCM Prompting

41 / 58

1. Analyse

Mine root causes

2. Compare

Benchmark trade-offs

3. Summarise

Synthesize memos

4. Extract

Pull specific dates/fees

5. Classify

Bucket by risk tier

6. Challenge

Stress-test plans

7. Prioritise

Rank by urgency

8. Simulate

Model disruption

9. Recommend

Multi-option choices

10. Identify

Flag uncertainties

Advanced Prompting

⛓️ Prompt Chaining: Step-by-Step Rigor

42 / 58

1

Extract

Pull raw data points

2

Organise

Structure into table

3

Analyse

Detect variances

4

Risk Audit

Spot failure modes

5

Options

Generate 3 solutions

6

Challenge

Devil's advocate test

7

Executive Memo

Final briefing draft

Cognitive Framework

🎯 The 3-Pass Method

43 / 58

PASS 1 — UNDERSTAND

"Summarise and structure the raw information."

Ensure AI accurately parses terms, timelines, and quantities without speculation.

PASS 2 — ANALYSE

"Identify patterns, root causes, and risks."

Detect operational anomalies, supplier dependencies, and cost drivers.

PASS 3 — DECIDE

"Provide 3 strategic options with trade-offs."

Formulate actionable recommendations with cost and lead time impacts.

Operational Cheat Sheet

🛠️ SCM Tool + Task Matrix

44 / 58

Daily Workflow Task	Best Starting Tool	Alternative Tool	Key Advantage
Drafting Stakeholder Emails	M365 Copilot (Outlook)	ChatGPT Plus	Direct in-app contextual drafting
Meeting Minutes & Actions	Teams Copilot	ChatGPT (Transcript)	Real-time speaker attribution
Large Spreadsheet Variance	Excel Copilot	ChatGPT Advanced Data	Python script execution on CSV
Carrier Contract Auditing	NotebookLM	Claude 3.5 Sonnet	Zero hallucination; citations linked to source
Geopolitical Disruption Tracking	Perplexity AI	Google Gemini 1.5	Live web search grounding with verified links

Productivity Transformation

🔄 Before & After AI Adoption

45 / 58

Traditional Logistics Work (Manual)	AI-Assisted Logistics Work (Accelerated)	Value Unlock
Read 50 long email threads	AI synthesizes context + generates action list	40 minutes saved / day
Search across folders for contract terms	Grounded AI extracts exact clause & liability cap	Instant SLA verification
Manually draft monthly management slides	AI generates first draft structured outline & tables	2 hours saved / report
Inspect 1,000 shipment spreadsheet rows	AI flags top 17 high-risk exceptions automatically	Zero-touch exception triage

Hyundai Glovis Strategy

🎯 10 High-Value Use Cases for Glovis

46 / 58

01

Shipment Exception Monitoring

02

Supplier Risk Scoring Radar

03

Automotive CKD Demand Forecast

04

PCC Maritime Route Optimization

05

Dynamic Warehouse DC Slotting

06

Geopolitical Chokepoint Alerting

07

Management Memo Automation

08

Carrier Contract & SLA Audit

09

Operations Meeting Intelligence

10

Autonomous Freight Audit Agents

Business Value

💰 Measuring True AI SCM ROI

47 / 58

TIME

Hours Saved

Reduction in manual reporting and data consolidation hours.

COST

Demurrage Avoided

Direct reduction in port dwell penalties and emergency air freight.

QUALITY

Zero Billing Errors

Elimination of transcription errors in customs and carrier invoices.

SPEED

Cycle Time

Order-to-dispatch response times accelerated by 35%.

REVENUE

Throughput Scale

Handling higher freight volumes without proportional headcount growth.

RISK

Early Warning

Disruptions spotted days before impact, preventing factory line stops.

Project Selection

📊 AI Use Case Prioritisation Rubric

48 / 58

Evaluation Factor	Core Scrutiny Question	High Priority Indicator
Business Impact	How much cost/time value is unlocked?	Direct impact on demurrage, fuel, or line continuity
Task Frequency	How often does this workflow recur?	Daily or continuous transactional execution
Manual Drag	How many manual hours are currently spent?	> 5 hours/week per team member
Data Availability	Is clean digital text/data accessible?	Structured ERP, AIS, or CSV log availability
Operational Risk	What is the downside if output is flawed?	Low/Manageable with a human approval gate

Daily Habit

⏱️ The 30-Minute Daily AI Workflow

49 / 58

1

5 min

Define task & exact constraints

2

5 min

Extract & clean input data

3

10 min

Execute structured prompt

4

5 min

Verify against primary facts

5

5 min

Finalize & take operational action

Interactive Simulation

🚨 Live Workshop: The Shipment Crisis Challenge

50 / 58

Disruption Scenario

A container vessel carrying critical EV inverter modules from Incheon to Chennai is delayed 6 days due to typhoon rerouting.

Assembly line stoppage projected in 72 hours
Air freight emergency quote: +$140,000
Alternative buffer stock located in Singapore 3PL

⏱️ Group Exercise: 7 Minutes

PROMPT PLAYBOOK

Act as a crisis logistics director. Assess: 1) Line-stop risk exposure, 2) Cost vs. transit trade-off between Singapore air-bridge vs. port expediting. Produce: A 1-page Executive Action Memo with Situation, Findings, 3 Ranked Options, and Recommended Next Steps.

Universal Framework

🌟 The Golden AI Formula

51 / 58

1

TASK

Clear business goal

2

DATA

Verified facts

3

PROMPT

7-part formula

4

AI

High-speed draft

5

VERIFY

Human audit

6

DECIDE

Leadership call

7

MEASURE

Track business ROI

Immediate Adoption

📅 7 AI Habits to Start Tomorrow

52 / 58

Morning & Operational Habits

1. Priority Brief: Start the morning with an AI summary of open dispatch tickets.
2. Actionize Threads: Turn long email chains into structured action tables.
3. Draft Outlines: Generate structured report drafts before writing narrative prose.
4. Devil's Advocate: Ask AI to challenge your logistics plans before management review.

Analytical & Scaling Habits

5. Anomaly Mining: Use AI data analysis on spreadsheets before manual row inspection.
6. Prompt Library: Save high-performing prompt templates in a shared team folder.
7. Agent Opportunity: Identify 1 repetitive task this month to turn into an AI agent flow.

Risk Safeguards

🚫 7 Things NOT to Do with AI

53 / 58

❌ Pitfalls (1 to 4)

1. Don't Blindly Trust: Never accept unverified dates or financial calculations.
2. Don't Leak Data: Never upload confidential pricing or customer PII to public tools.
3. Don't Accept Fabrications: If a citation or rate looks unusual, verify with source.
4. Don't Skip Human Gate: Never let AI directly dispatch orders without oversight.

❌ Pitfalls (5 to 7)

5. Don't Automate Chaos: Fix broken manual processes before applying AI agents.
6. Don't Use Wrong Tools: Don't use conversational chat for large tabular parsing.
7. Don't Measure Vanity: Measure dollars saved and risk reduced, not prompt count.

Self-Assessment

📘 Your Personal AI Workflow Playbook

54 / 58

For every recurring operational bottleneck, ask these 6 filter questions:

1. Is it repetitive?

Occurs on a predictable daily or weekly cycle.

2. Is it data-heavy?

Requires processing large volumes of text/numbers.

3. Is data digital?

Information exists in accessible digital text/CSVs.

4. Can AI draft it?

LLM can synthesize, calculate, or format.

5. Can you verify it?

Clear ground truth exists to audit accuracy.

6. Is value high?

Saves >2 hours or directly mitigates risk.

Strategic Mindset

🎯 The Core Mindset Shift

55 / 58

The Flawed Question

"What can ChatGPT do?"

Leads to aimless experimentation and superficial tool usage without business ROI.

The Executive Question

"Where does work slow us down?"

Identifies exact operational friction, broken transload handoffs, and reporting delays.

1. Identify Process

2. Attach Data

3. Choose AI

4. Precision Prompt

5. Human Verify

6. Scale Impact

The Future of Logistics

NOT HUMAN vs AI —
HUMANS + AI

Human Logistics Expertise + AI Intelligence + Enterprise Data + Automated Workflows
= SMARTER, RESILIENT SUPPLY CHAINS

"The biggest AI advantage will not come from knowing the most tools. It will come from redesigning the way we solve supply chain challenges."

Learn AI Apply AI Verify AI Scale AI

Take-Home Handout

📑 Trainer's Live Prompt Cheat Sheet

57 / 58

01 // ANALYSE & COMPARE

"Analyse this logistics performance data and identify the top 5 operational bottlenecks." "Compare these 2 carrier proposals across total cost, transit SLA, and fuel risk."

02 // EXTRACT & SUMMARISE

"Extract all demurrage charge terms, grace days, and payment deadlines into a table." "Summarise this 20-page port congestion report for senior leadership in 5 bullets."

03 // CHALLENGE & PRIORITISE

"Act as a devil's advocate and highlight 3 hidden failure points in this route plan." "Rank these 10 shipment exceptions by OEM production impact and urgency."

04 // EXECUTIVE BRIEF

"Convert this analysis into a 1-page executive brief: Situation, Findings, 3 Options, and Next Steps."

Facilitator Guide

⏱️ Recommended 2-Hour Delivery Agenda

58 / 58

Time (Min)	Module Section	Core Content Covered	Practical Interactive Component
0–10 min	AI & Smart SCM	Evolution from Copilots to Agents	Hyundai Glovis operational opportunity map
10–25 min	AI Tools Matrix	ChatGPT, Gemini, Claude & Copilot	5-Second tool selection decision exercise
25–45 min	Prompt Engineering	7-Part formula + 7 power techniques	Live prompt refactoring (Weak ➔ Excellent)
45–70 min	AI for Logistics	Demand, shipping, routes & DC slotting	Busan-Chennai shipment re-routing case
70–90 min	Risk & Decisions	Supplier risk scoring & scenario planning	Multi-variable disruption trade-off matrix
90–105 min	AI Agents	Digital workforce & autonomous loops	Exception triage agent workflow breakdown
105–115 min	Responsible AI	Confidential data & verification loop	Data hygiene checklist & hallucination audit
115–120 min	Live Simulation	Shipment Crisis Resolution Challenge	Group crisis memo synthesis & wrap-up

AI-POWERED PRODUCTIVITY & SMART SUPPLY CHAIN

🗺️ Today’s Learning Roadmap

AI & GenAI

AI Tools

Prompting

Productivity

Smart SCM

Decisions

AI Agents

Responsible AI

🏢 Why AI Matters to Hyundai Glovis

Integrated Glovis Operational Footprint

📈 The AI Evolution in Enterprise

🤖 What is Artificial Intelligence?

Core Definition

⚠️ Non-Negotiable Operational Reality

✨ What is Generative AI?

Multimodal Creative & Analytical Capabilities

The Iterative Generation Loop

Ask

Gen

Review

Tune

🗺️ Enterprise AI Tool Map

⚖️ Model Matrix: ChatGPT vs Gemini vs Claude vs Copilot

⏱️ The 5-Second Tool Selection Rule

Draft / Brainstorm / Code

Inside Office Apps

Live Web Intelligence

Multi-Document Binders

🎯 Prompt Engineering: Quality In = Quality Out

❌ Weak Prompt

⚠️ Better Prompt

✅ Production-Ready Prompt

📐 The 7-Part Prompt Formula

Role

Context

Task

Data

Constraint

Output

Verify

📋 Master Prompt Template

💡 Technique #1: Ask AI to Ask Questions First

Traditional Flawed Approach

🎭 Technique #2: Assign Precise Professional Roles

Why Role Priming Changes Output:

📊 Technique #3: Enforce Structured Tabular Outputs

Output Schema Preview

⚔️ Technique #4: Ask AI to Challenge You

🔍 Technique #5: Force AI to Show Uncertainty

Why This Prevents Disasters:

🔀 Technique #6: Request Multi-Option Trade-Offs

Don't Ask:

👓 Technique #7: Use AI as a Second Reader

High-Value Application Areas:

✉️ AI for Email Communication

Everyday Email Workflows:

🎙️ AI for Meeting Intelligence

Generated Action Matrix

📑 AI for Management Reports

📊 AI for Excel & Spreadsheet Analytics

Complex Formulas

Data Cleansing

Variance Detection

Python Automation

🚚 Logistics AI Use-Case Ecosystem

📈 AI for Automotive Demand Forecasting

Forecasting Variables Combined:

🛡️ AI for Supplier Risk Intelligence

Simulated Supplier Risk Card

🚛 AI for Transport & Route Optimization

🎯 Simulated AI Decision Artifact

🏬 AI for Warehouse & Inventory Optimization

Core DC Impact Metrics:

🚢 AI for Maritime & Ocean Logistics

Maritime Factors Synthesized:

🎯 AI for Exception Triage & Filtering

Operational Paradigm Shift:

🌐 AI for Geopolitical & Trade Disruption

AI-POWERED PRODUCTIVITY &
SMART SUPPLY CHAIN