Wonderful experience presenting Tata Steel UK Supply Chain's Robotic Process Automation (RPA) and Artificial Intelligence (AI) journey in the international conference on "RPA and AI in Customer Facing Supply Chain" in Amsterdam, May 2018. Pre-presentation interview link: https://events.marcusevans-events.com/cs122-interview-tata_steel/
3. Tata Steel 3Slide
Tata Steel – global operations and commercial presence
3
Western
Europe
Steel making operations
Distribution and
downstream assets
North America
Western
Europe
Scandinavia
CIS
Western
Africa
South
Africa
Latin
America
China
Japan
India
Hong Kong
Sales offices
SE Asia
Key
Turkey
CEE
4. Tata Steel 4Slide
Extended Trans-European Supply Chain challenges system integration 4
French
Downstream
Asset
UK Manufacturing Hub
UK Port
French port
Spanish port
French barge
Haulier Transport
Spanish
Downstream
Asset
Further
processing and
supply to end
customer
Further
processing and
supply to end
customer
New order placement / stock replenishment
New order placement / stock replenishment
Coil Stock
Coil Stock
5. Tata Steel 5Slide
TSE Vision for Digital Enterprise and AI Adoption in Supply Chain 5
Base Layer
(Disparate Process + Multiple ERPs)
Data Integration and
Visualization
(Single Source of truth)
Automation
(RPA & ERP
enhancements)
Big data
Analytics
AI
Predictive
Intelligent
Smart
Real Time Digital
Steel Enterprise
• Dark/Hidden Spot using NLP/
Machine Learning / Deep learning
• Collusive behavior / Process
consistency
• Intelligent Command Centre
• Uses Cognitive using external
factors of weather, natural
disasters
• Analytics model
• Process Modelling
• Leverage Big data
• Influence Business
Outcomes
• Order Entry Automation /RPA
• Order Complexity /MTA &
MTO
• Critical trigger for Order to
delivery
• Data arching layer interface
to ERPs
• Unified view and dashboard
of Steel
• single data model and
single user interface
• Multiple and Disparate
systems
• Data and ERP Fragmentation
• Low Process Maturity
• Steel Industry
Characterization
Unified view and
visibility into steel
availability
Cognitive
Enhanced End
Customer
Experience
Predictable Operations
Intelligent Delivery
consistency
Real Time Visibility
No visibility to
steel availability,
Not sure of
Timely Order
delivery
Drive Analytics for
Inventory
Process modelling
Seamless Delivery
Cater to dynamic
order pattern and
consumption
behavior
6. Tata Steel 6Slide
RAPID – Tata Steel’s Stock integration platform across disparate MRP landscape
• Automatic periodic run to provide a refreshed and updated
view of the end stock
• Presents a visual tool by extracting relevant data from
disparate manufacturing systems and SAP
• Aids stock replenishment by providing a visibility of current
stock and end-consumption
7. Tata Steel 7Slide
However, with numerous SKUs, traditional order entry process poses a
challenge
7
Creation of Purchase orders on SAPCreation of Purchase orders on SAP
Customer Purchase OrderCustomer Purchase Order
Staging FileStaging File
SAP Order Entry FileSAP Order Entry File
Order AcknowledgementOrder Acknowledgement
OCR Process
Conversion to structured data
Manual typing of orders (15mins-20mins)
Successful order creation manually
Slow
process
Slow
process
Prone to
human
errors, often
leading to
undesired
production
Prone to
human
errors, often
leading to
undesired
production
Repetitive activity
is not always
considered job
enrichment
Repetitive activity
is not always
considered job
enrichment
8. Tata Steel 8Slide
Robotic Process Automation at work - TCS Zeus for order entry 8
• Entirely mimics human behaviour by automatically typing all
fields onto SAP and submitting the order
• Intelligent to determine appropriate order quantity based on
Minimum Order Quantity (MOQ) restrictions on SAP and goes
for the closest match
• On encountering any error, Zeus enters the error into a log and
moves ahead to complete other order entries
Zeus Recorded Demo
9. Tata Steel 9Slide
Zeus Benefits 9
• No change in working of the core applications like
SAP. Zeus acts like a human that only interacts
with SAP
• Intelligently designed to take judgements on
particular conditions like coil weight and week
when capacity is available to place order
• Handles errors well by recording them into a log
and progressing to other orders
10. Tata Steel 10Slide
Big data storage and capture through RAPID system 1
0
Order Fulfilment Suite
Health of stock measurementOrder progress through manufacturing hub
11. Tata Steel 11Slide
Big data analysis and actionable insights – stock out and overstock prediction1
1
Future stock health prediction
Stock
out
Low
stock
Adequate
stock
Over
stock
Marginally
High stock
+8 weeks stock prediction
12. Tata Steel 12Slide
Big data analysis and actionable insights – customer consumption behaviour1
2
Enhanced customer
centricity through joint
collaboration with
downstream units and
customers to drive
improved service
performance / optimized
inventory across the value
chain
Stock consumption analysis and insights
Daily customer
consumption
Smart inventory
management?
13. Tata Steel 13Slide
The Future: Machine learning and cognitive intelligence – Predicting
Variabilities to Planned Exit Date and Planned Delivery Date
1
3
Customer Order Planned Exit Date Planned Delivery Date
Common cause
plant variability
Changing customer
priorities
Special cause
plant variability
Port
Strike
River
Flooding
Snowfall
~ ~ ~ ~ ~ ~
Variability Variability
14. Tata Steel 14Slide
IgnioTM – A Comprehensive Intelligent Suite for ‘Order to Delivery’ 1
4
Sense
Think
Act
Detect patterns – Customer Order, SKU
footprint, Order exceptions,
Prediction Model – Inventory
Optimization Planned Exit/Adherence to
Delivery
Triage and Recommend – Command
Center, Prediction model evaluation
API and RPA – Order entry automation,
Triggers Planning, Work scheduling /
prioritization model, Gamification
Learn
Cognitive – predictable, touchless,
Intelligent, Machine learning all of which
leading to improved business outcomes
Intelligent & Predictive
delivery for customers
Decipher Insights,
Proactive triggered
actions
Insights to Interception
Improved ‘On Time
Delivery’
Optimized ‘Inventory’
15. Tata Steel 15Slide
Resolving Planned Exit Date (PED) variance through Ignio – way ahead 1
5
Current tracking - reactive
Future tracking – predictive
Customer name =“Renault”; product ordered = “sheet metal”, Order month
=November; PED accuracy =90%
Customer name
=“Renault”
Product =“sheet
metal”
Product =“rolled
steel”
Order Month
=“November”
Order Month
=“january”
PED accuracy=“90%”
16. Tata Steel 16Slide
What the future holds – real time prediction of Over/understock intelligently1
6
*Tata Steel presently uses its enhanced in-house RAPID system capabilities to predict overstock /
understock
17. Tata Steel 17Slide
What the future holds - predicting Overstock/Understock intelligently 1
7
~ 50 K Ton new Orders every
week
2Hub &
8 Downstream
4 Production
Strategies
50 +
Warehouses
5
Customer
Streams400 K Ton Avg Stock
managed at any given point
32 diff activities performed to
keep the stock moving daily
2200 + Stocked Unit
Excessive Stock built (~3x of
avg requirement) post migration
into new platform
High dwell of SKUs in Over-stock
and Out-of-stock - >21 Days
against a tolerance of 7 Days
Re-active approach – distressed SKUs
~ 17% against tolerance of 5%
Successful POCwith proven
measures
3 Week ahead
Preventive recommendations
> 2Mn GBP potential
BVA
D
C
O
E
Order
Entry
Decoupling
Point
Notional
Slab
Hot Roll
Pickling
Cold Roll
Annealing
Quality
Check
Packaging
Stock
Transfer
Galvanizing
18. Tata Steel 18Slide
1
8
The future world of AI and cognitive systems….
Customer Order Planned Exit Date Planned Delivery Date
Common cause
plant variability
Changing customer
priorities
Special cause
plant variability
Port
Strike
River
Flooding
Snowfall
~ ~ ~ ~ ~ ~
Variability Variability
MitigationsandCognitiveLearning
MitigationsandCognitiveLearning
Mitigations and Cognitive Learning