MANAGEMENT OF DIGITAL BUSINESS SYSTEMS
CIA-3
Priyanshu Chhabariya
23221072
Christ University School of Business and Management
Masters of Business Administration
Dr. Durgansh Sharma
1 OCTOBER 2023
ARTIFICIAL INTELLIGENCE FOR
IMPROVING THE PROCUREMENT
EXPERIENCE OF NON-STOCK ITEMS AT
INDIAN RAILWAYS
SUMMARY OF THE CASE STUDY
In the case study, Indian Railways (IR) faces challenges in
efficiently procuring non-stock items, prompting Chief Technology
Officer (CTO) Sumana G. to explore an artificial intelligence (AI)
solution. With IR's vast network and inventory, keyword-based
searches in their integrated monitoring system often led to
incorrect or inefficient results. This mismatch problem resulted in
lengthy vendor on boarding procedures and wasted time. Sumana
enlisted the help of Reddy. Reddy, an AI expert, proposed to use
semantic similarity matching to address this problem. He used a
clustering algorithm to group similar items together based on their
meaning, rather than just their keywords. Eddy tested the
algorithm with a small dataset of 197 item descriptions and
achieved an accuracy of 85%. He was confident that he could
improve the accuracy with a larger dataset. IR provided Reddy
with a dataset of 4,000 data points to test the algorithm with.
Reddy is currently refining the algorithm to address scalability,
accuracy, and timeliness concerns. If Reddy is successful, the AI-
based solution will help IR to improve the efficiency and
effectiveness of its procurement process.
Q1. Explain how AI helps improve the procurement
experience of non-stock items in Indian railways?
ANS 1- Artificial Intelligence (AI) can significantly improve the
procurement experience of non-stock items in Indian Railways in several
ways:
Improve search efficiency:
AI-powered search engines can understand the semantics of search
queries. Unlike traditional keyword-based searches, AI algorithms can
identify items based on their meaning, providing more accurate and
relevant search results. This reduces the probability of failed searches
and minimizes the time field agents have to spend looking for specific
items.
Match semantic similarity:
AI can group items based on their semantic similarity, meaning it can
group together items with similar characteristics and descriptions. This
helps field personnel not only find the exact item they are looking for, but
also find similar alternatives that may meet their needs.
Improve data standardization:
AI can help standardize item descriptions and nomenclature. By
analyzing historical supply data, AI algorithms can recommend
standardized naming conventions, reducing discrepancies in the
description of out-of-stock items, helping to recover orders The previous
product was easier.
Quick access to historical data:
AI can give field staff access to historical shopping data for items that are
not in stock. This allows them to quickly review past orders, prices and
suppliers, helping them make informed decisions and negotiate prices
more easily.
Reduce excess effort:
With advanced search capabilities, AI can help avoid redundant supplier
on boarding procedures. Field agents will not initiate the supplier
registration process if the requested items are already in the system,
saving time and resources.
Streamlined Procurement Process:
AI can automate aspects of the procurement process, such as supplier
selection and price comparison. This streamlines the overall procurement
process, making it more efficient and reducing human intervention.
Continuous improvement:
AI algorithms can adapt and learn from user interaction and feedback. As
more data is processed, AI systems can continuously improve search
and recommendation capabilities, further improving the shopping
experience over time.
Cutting the cost:
By optimizing the procurement process, AI can help Indian Railways
negotiate better rates with suppliers, identify cost reduction
opportunities, and minimize time and resources spent on research
inefficiencies and redundant supplier integration.
Q2. How would the Data Lakes and Data Hubs help design a
system like Blinkit or Zepto?
ANS 2- Data Lakes and Data Hubs are important components in
designing and implementing systems like Blinkit or Zepto, which are
typically delivery and logistics platforms. These platforms require
efficient data management, processing, and analysis to optimize
operations and enhance customer experiences.
Data Lakes allow for the storage of vast amounts of structured and
unstructured data at a lower cost. In the context of delivery and logistics
platforms, this can include order data, customer information, GPS tracking
data, warehouse inventory, and more. Data Lakes enable the centralized
storage of data from various sources, making it accessible for analysis and
processing.
Data Hubs facilitate real-time data ingestion from multiple sources. For
platforms like Blinkit or Zepto, this means continuously collecting data
from various touch points such as mobile apps, websites, IoT devices, and
delivery vehicles. Real-time data ingestion ensures that the platform
operates with the most up-to-date information, allowing for accurate
tracking and monitoring.
Data Lakes enable the integration of data from different departments and
systems within the organization. For delivery and logistics platforms, this
means combining data from customer service, order processing, inventory
management, and delivery tracking. Data Lakes serve as a unified
repository for these disparate data sources.
Both Data Lakes and Data Hubs play a crucial role in enabling machine
learning and AI applications. These platforms can feed data into machine
learning models to optimize delivery routes, predict demand, detect
anomalies, and provide personalized recommendations to customers.
Data Lakes and Data Hubs play a foundational role in the design and
operation of delivery and logistics platforms like Blinkit or Zepto. They
enable the collection, storage, integration, processing, and analysis of data,
which is essential for optimizing operations, improving customer
experiences, and scaling the platform as it grows.
Q3. Discuss the usage of Advanced Augmented Reality,
Wearable Technologies, and Digital Twins, which would be
helpful in real-time troubleshooting during manufacturing
operations.
ANS 3- The integration of Advanced Augmented Reality (AR), Wearable
Technologies, and Digital Twins can greatly enhance real-time
troubleshooting during manufacturing operations by providing a holistic
and data-driven approach.
1. Advanced Augmented Reality (AR):
➢ Remote Assistance: AR enables off-site professionals to assist on-
site technicians with real-time instruction. Technicians can receive
visual instructions, notes, and 3D models projected onto real
equipment using AR headsets or mobile devices, assisting in
maintenance.
➢ Digital Overlays: AR may apply digital information onto real
machinery, such as blueprints, service instructions, or real-time data
from sensors. This gives technicians with quick context, allowing
them to diagnose difficulties and perform duties more effectively.
➢ Interactive Training: AR can be utilized for interactive training as
well as on-the-job learning. Technicians can use AR devices to access
training modules and simulations, allowing them to learn new skills
and troubleshoot problems more efficiently.
2. Wearable Technologies:
➢ Smart Glasses: Wearable smart glasses equipped with cameras,
sensors, and displays can provide hands-free access to critical
information. Technicians can view schematics, checklists, and real-
time data without needing to use their hands, allowing them to focus
on troubleshooting tasks.
➢ Wearable Sensors: Wearable sensors can monitor vital signs of
technicians, ensuring their safety during troubleshooting activities.
They can also track environmental conditions, equipment vibrations,
and other parameters that might be relevant to identifying issues.
➢ Voice Commands: Wearable devices with voice recognition
capabilities enable technicians to issue commands, access
information, and record observations without the need for manual
data entry. This can enhance efficiency and reduce errors.
3. Digital Twins:
➢ Real-time Simulation: Digital Twins create virtual replicas of
physical assets or manufacturing processes. By coupling these digital
replicas with real-time data from IoT sensors and other sources,
technicians can visualize and analyze the actual state of equipment in
real-time, helping them identify abnormalities and troubleshoot
issues.
➢ Predictive Maintenance: Digital Twins can predict when equipment
is likely to fail or require maintenance based on real-time data. This
proactive approach allows technicians to address potential issues
before they cause production disruptions.
➢ Collaborative Problem Solving: Digital Twins can be accessed and
manipulated by multiple stakeholders simultaneously. Technicians,
engineers, and experts from various locations can collaborate on
troubleshooting efforts, leveraging a shared digital representation of
the equipment.
