Digital Transformation
• Digitaltransformation is the process of using digital technologies to
create new — or modify existing — business processes, culture, and
customer experiences to meet changing business and market
requirements.
• ERP, CRM, SCM implementation in organizations
5.
For Example
• Workflowof travel application for a staff in IIM Indore
• Earlier-
Employee Supervisor Department Head Dean Director
• After ERP-
Employee Department Head Dean Director
6.
• Inaccurate requirements
•Uninvolved top management
• Shifting project objectives
• Inaccurate estimates
• Unexpected risks
• Dependency delays
• Not enough resources
• Poor project management
Possible Reasons for large IT Project Failure
Market Basket Analysis
•Market Basket Analysis is one of the key techniques used by large retailers to
uncover associations between items
• It works by looking for combinations of items that occur together frequently in
transactions
• For example:
1) Bread and Butter are bought together, so they are placed near to each other
2) Toothbrush and Toothpaste
3) Milk, cheese, curd
Similar examples in Banking/ Insurance Sector?
11.
Technological Enablers ofDigital
Transformation
• Internet of Things (IoT) :
• IoT refers to the network of interconnected devices and objects that can collect and
exchange data.
• IoT is applied through telematics, where sensors and devices gather data from vehicles,
homes, or wearable devices.
• Artificial Intelligence (AI) and Machine Learning:
• AI involves the creation of machines capable of simulating human intelligence, including
learning and problem-solving.
• Machine learning, a subset of AI, involves algorithms that improve performance over time
through data analysis.
• AI and machine learning are transforming the banking industry by doing risk assessment,
fraud detections and so on.
12.
Machine Learning ClassificationExample
S.No Age Gender Income Loan Default
1 32 Male 10 lpa No
2 24 Female 12 lpa Yes
3 56 Male 19 lpa No
4 23 Male 8 lpa Yes
5 22 Female 15 lpa Yes
6 45 Male 12 lpa No
7 51 Female 8 lpa No
8 28 Female 10 lpa ?
9 18 Male 4 lpa ?
10 72 Male 28 lpa ?
12
Image, Audio andVideo generation
Andrew Ng
Voice generation
https://voicemaker.in/voice-samples
28.
Image, Audio andVideo generation
Video generation
https://www.youtube.com/watch?v=HK6y8DAPN_0&ab_channel=OpenAI
30.
AI is aset of tools
Unsupervised
learning
Generative
AI
Supervised learning
(labeling things)
AI
Reinforcement
learning
31.
Supervised learning (labelingthings)
Input (A) Output (B) Application
Email
Ad, user info
Image, radar
info X-ray
image Image of
phone Audio
recording
Restaurant
reviews
Spam?
(0/1) Click?
(0/1)
Position of
other cars
Diagnosis
Defect? (0/1)
Text transcript
Sentiment
Spam filtering
Online advertising
Self-driving car
Healthcare
Tools like Cashify
Speech recognition
Reputation monitoring
32.
Generating text usingLarge Language Models (LLMs)
Text generation process
I love eating
prompt
food
my _________
snacks
AI output
33.
How Large LanguageModels (LLMs) work
My favorite food is a
My favorite food is a bagel
My favorite food is a bagel with
My favorite food is a bagel with cream
bagel
with
cream
cheese
LLMs are built by using supervised learning (A→B) to repeatedly predict the next word.
My favorite food is a bagel with cream cheese
Input (A) Output (B)
When we train a very large AI system on a lot of data (hundreds of billions of words),
we get a Large Language Model like ChatGPT.
34.
Bias and Toxicity-Gen AI
An LLM can reflect the biases that exist in the text it learned from.
Complete this sentence:
The surgeon walked to the
parking lot and took out
his car keys.
Complete this sentence:
The nurse walked to the
parking lot and took out
her phone.
assumed male assum ed fem ale
Some LLMs can output toxic or other harmful speech, but
most models have gotten much safer over time.
Andrew
Ng
35.
Small Case Studies-1
•At a bank, information about Mr. Naveen
• He has three accounts: savings, credit card, and a car loan
• He makes five deposits and 25 withdrawals per month
• He never visits a branch in person
• He has a total of Rs 3,50,000 in cash deposited
• He owes a total of Rs 1,15,000 between his credit card and car loan
•What is the next best offer you can place to him in an e-mail?
36.
Example (Contd)
• IfMr. Naveen’s web behavior is examined and we got additional
information
• He browsed home loan rates five times in past month
• He viewed websites like makaan.com, housing.com, 99acres.com
• He explored home loan options (i.e., fixed versus variable, 15- versus 30-year)
twice in the past month
•Then, it is pretty easy to decide what to discuss next with Mr. Naveen
37.
Privacy Issues inSocial Network
• Facebook/ Instagram
• Whatsapp
• Concept of privacy calculus- Tradeoff
• Classic case of Truecaller
38.
Privacy Issues inLocation Data
• Retail store example- wifi- geofencing
• Fastag removal on highways
How to avoidthe hackers to gain unauthorized access to the systems?
• Create complex passwords & change them often
• Install anti-virus programs and highly secured firewall
• Log out of accounts when you are done with them
• Make sure you are on the official website when entering passwords
• Use two-factor authentication
• Clear your browser cookies and never ever save your passwords in the
browser and so on
42.
Digital Personal DataProtection (DPDP) Act
2023
• The Act deals with safeguards around personal data only and keep
non-personal data (information that does not reveal the identity of an
individual) out of its ambit
• It will continue to evolve (with the changing technology)
• Entities that fail to take “reasonable security safeguards” to prevent
personal data breaches will be fined as high as Rs 250 crore.
43.
Digital Personal DataProtection (DPDP) Act
2023
• The draft data protection bill 2022 states that-
• “When it comes to data localization rules, where the government
mandated all companies dealing with sensitive data of Indian users to
keep a copy within its borders. The government has allowed the
transfer of data and its storage in “trusted geographies” in the revised
draft of the data protection Bill, doing away with the data localisation
requirement proposed in the earlier version. The government will
define which geographies are “trusted” from time to time”
• The clause is now removed in DPDP 2023