Business Analytics

Module-1

An Overview of Business Intelligence, Analytics, Data Science, and AI:

Lecture Notes: Changing Business Environments and Evolving Needs for Decision
Support and Analytics

1. Introduction to Decision Support and Analytics

• Modern organizations operate in fast-changing environments filled with

uncertainty.

• Decision making is increasingly data-driven, leveraging analytics and IT

systems.

• The traditional scope of IT (payroll, bookkeeping) has now expanded to strategic

and managerial functions.

2. Importance of Decision Making in Organizations

• Decision making is central to organizational success.

• It is becoming more complex due to internal dynamics and external pressures.

• Poor decisions can be costly, while good decisions provide competitive

advantage.

Types of Organizational Decisions (De Smet et al., 2017):

1. Big-bet decisions – high-risk, often strategic.

2. Cross-cutting decisions – repetitive, high-stakes, and collaborative.

3. Ad hoc decisions – infrequent and situation-specific.

4. Delegated decisions – routine and handled by individuals or small teams.

3. The Decision-Making Process

• Historically seen as an art form, based on:

o Experience

o Intuition

o Judgment

o Creativity
 • Shift toward scientific and analytical approaches.

• Research shows methodical, persistent work outperforms charisma or

communication alone in management.

4. Role of Analytics in Decision Making

• Analytics enables:

o Descriptive reports (what is happening)

o Predictive modeling (what could happen)

o Prescriptive decisions (what should be done)

• Massive data volumes (doubling every ~2 years) now power these insights.

5. Technologies Enabling Decision Support

a. Group Communication & Collaboration

• Remote and distributed teams use:

o Smartphones

o Collaboration platforms

• Supply chain integration needs real-time data sharing.

b. Improved Data Management

• Diverse, multimedia data stored across locations.

• Systems ensure secure, fast, and integrated access.

c. Big Data & Data Warehousing

• Technologies include:

o Hadoop/Spark

o Parallel processing

o Cloud storage

• Supports analysis of structured and unstructured data at scale.

d. Analytical Support

• Tools help evaluate alternatives, perform simulations, and predict outcomes.

 • Decision-making is more efficient, cost-effective, and evidence-based.

e. Overcoming Cognitive Limits

• As per Simon (1977): human minds have limited capacity to process complex
info.

• Computers extend human cognition by handling vast information quickly.

f. Knowledge Management Systems (KMS)

• Collect formal and informal knowledge from operations.

• Text analytics and AI help extract insights from communication patterns.

g. Anywhere, Anytime Decision Support

• Mobile and wireless tech allow decision makers to:

o Access data

o Analyze in real-time

o Collaborate instantly

• Decision cycles are shorter and more responsive.

6. Innovation and Artificial Intelligence (AI)

• AI brings intelligence and automation to the decision-making process.

• AI enhances:

o Data analysis

o Pattern recognition

o Decision recommendation

• Synergy between AI + Analytics improves:

o Accuracy

o Speed

o Strategic insight
 7. Conclusion

• Decision-making in modern businesses is data-intensive and technology-

supported.

• Organizations must embrace:

o Analytical thinking

o Advanced technologies

o Collaborative decision-making

• Tools such as BI, analytics, and AI are indispensable for managerial success.

Lecture Notes: Decision-Making Processes and Computerized Decision Support

Framework

1. How Can Computers Help Overcome Cognitive Limits of Humans?

• Cognitive Limits refer to the limited ability of the human brain to process, store,
and recall vast or complex information (Simon, 1977).

• Computers assist in overcoming these limits by:

o Quickly storing and retrieving large datasets.

o Performing complex calculations without error.

o Providing analytical models and simulations to evaluate decisions.

o Reducing information overload through visualization tools (e.g.,

dashboards).

o Supporting recall and documentation of previous decisions and their

outcomes.

o Using AI and machine learning to detect patterns humans may overlook.

 2. Simon’s Classical Decision-Making Model

Herbert A. Simon proposed a structured, rational model comprising four main phases:

a. Intelligence Phase (Problem Discovery)

• Examine and understand the real-world situation.

• Activities:

o Define goals and objectives.

o Identify problems/symptoms.

o Collect relevant data.

o Classify and decompose the problem.

o Assign problem ownership (responsibility).

• Example: Sales drop → Is it the real problem or a symptom (e.g., poor customer
service)?

b. Design Phase (Model Development)

• Develop models and define alternatives.

• Key tasks:

o Abstract and simplify reality using assumptions.

 o Identify variables and their relationships.

o Define criteria for evaluating alternatives.

o Test and validate the model.

• Models: Simplified representations of reality used to evaluate solutions.

c. Choice Phase (Solution Selection)

• Select the best course of action among alternatives.

• Key tools:

o What-if analysis: test parameter changes.

o Goal-seeking: adjust inputs to meet desired outcomes.

o Sensitivity analysis: assess robustness of solution.

• May return to the Design phase if chosen alternative is inadequate.

d. Implementation Phase (Putting Solution into Action)

• Enforce and execute the chosen solution.

• Involves:

o Change management and user training.

o Stakeholder communication and expectation setting.

o Project management and performance monitoring.

o Data collection for feedback and continuous improvement.

• Challenges: Resistance to change, vague boundaries of implementation, long-

term follow-up.

3. Monitoring/Feedback

• Sometimes treated as a fifth phase.

• Refers to tracking performance post-implementation.

• Aims to:

o Assess success/failure of decisions.

o Provide insights for future decision-making.

o Validate accuracy of models and assumptions used earlier.

 4. Key Concepts in Decision Support

a. Problem Classification

• Categorize problems based on structure:

o Structured: Routine, repetitive (e.g., reorder inventory).

o Unstructured: Novel, vague (e.g., entering new market).

o Semi-structured: Contains both elements.

b. Problem Decomposition

• Break down a complex problem into subproblems.

• Often, subproblems are easier to solve and provide insights into the main issue.

c. Problem Ownership

• Assign responsibility to individuals/groups.

• A problem can only be solved when it is “owned” by someone with authority and
ability.

5. Issues in Data Collection and Estimation

Decision-making quality depends on data quality. Common challenges:

• Lack of available or accurate data.

• High cost of data acquisition.

• Subjective estimates.

• Data security concerns.

• Overload of irrelevant or excess data.

• Outcomes may span over time → use of present-value analysis.

• Historical data may not match future trends.

6. Case Study: Analytics in Action – Making Elevators Faster

• Problem: Complaints about slow elevators.

• Initial solutions: Add elevators, operators, etc.

 • Real problem: Perceived waiting time.

• Final solution: Add mirrors near elevators.

• Lesson: Identifying the correct problem leads to simpler, cost-effective

solutions.

7. Summary

• Decision making is a systematic, iterative process.

• Simon's model provides a structured framework.

• Computerized Decision Support Systems (DSS) assist by:

o Managing vast data.

o Supporting all phases of decision making.

o Providing tools like simulations, models, and analytics.

• The intelligence phase is crucial—correct problem identification determines

solution quality.

• Implementation and feedback ensure decisions are actionable and improve

over time.

Lecture Notes: Evolution of Computerized Decision Support to Analytics/Data

Science

1. Introduction

• The evolution of analytics spans from routine reporting systems in the 1970s to
modern AI-driven, Big Data-powered decision support systems.

• Drivers of change:

o Increasing data volumes

o Advances in hardware & software

o New data sources (IoT, social media, sensors)

o Business need for timely & accurate decisions

2. Early Stages: 1970s – Decision Support Systems (DSS)

• Management Information Systems (MIS): Produced structured, periodic

reports.

• Purpose: Inform managers of past events (daily, weekly, monthly).

• Limitations: Static reporting, lack of flexibility, not real-time.

• Scott-Morton (1971) definition of DSS:

o “Interactive computer-based systems to help decision makers use data &

models for unstructured problems.”

• Keen & Scott-Morton (1978):

o Coupled human expertise with computer capabilities.

o Addressed semi-structured decision problems.

• Operations Research (OR):

o Mathematical models, optimization, and heuristics for constrained

resources.

3. Late 1970s–1980s: Expert Systems & Enterprise Integration

• Expert Systems (ES):

o Captured expert knowledge as if–then rules.

o Made expertise available on-demand.

• Enterprise Resource Planning (ERP):

 o Integrated departmental systems into a unified framework.

o Enabled single version of truth.

• Relational Database Management Systems (RDBMS):

o Improved data storage, relationships, and reduced redundancy.

• On-demand Reporting: Shift from fixed periodic reports to flexible, as-needed

reports.

4. 1990s: Executive Information Systems & Data Warehousing

• Executive Information Systems (EIS):

o Dashboards, scorecards, and visual reporting for executives.

• Data Warehousing (DW):

o Middle-tier repository for enterprise-wide decision making.

o Isolated reporting from transactional systems (ERP).

• Benefits:

o Maintained transaction integrity.

o Enabled versatile and timely decision-making.

5. 2000s: Business Intelligence (BI) Era

• DWs evolved into BI systems:

o Real-time and right-time data warehousing.

• Data Mining / Text Mining:

o Extracting new insights from large data volumes.

• Service-Oriented Architecture (SOA) & Software-as-a-Service (SaaS):

o Affordable analytics for SMEs.

o Pay-as-you-use model for software and infrastructure.

6. 2010s: Big Data & Social Media Analytics

• New data sources: IoT, RFID tags, smart devices, clickstream logs, wearables.
 • Challenges:

o Unstructured data

o Volume, Variety, Velocity (3Vs of Big Data)

• Technologies:

o Hadoop, MapReduce, NoSQL databases.

• Advanced Analytics:

o AI & Deep Learning for image, voice, and gesture recognition.

• User Interfaces:

o Voice assistants (Alexa, Google Home)

o Gesture-based controls

7. 2020s and Beyond

• Streaming Analytics:

o Real-time decision support using IoT sensor data.

• Cloud-Based Data & Processing:

o Massive scalability and global accessibility.

• Generative AI & Chatbots (e.g., ChatGPT):

o Conversational and contextual analytics.

• Future Trends:

o Faster paradigm shifts in analytics.

o Wider accessibility of data-driven insights.

8. Impact of COVID-19 on Analytics

• Positive:

o Highlighted importance of data-driven decision-making (e.g., disease

spread modeling).

• Negative:

o Short-term reduction in analytics investments due to economic

slowdown.
 • Prediction:

o Focus on analytics will return post-economic recovery.

9. Key Takeaways

• Analytics evolved from basic reporting to AI-driven real-time decision

support.

• Each decade brought technological and methodological innovations.

• The pace of evolution is accelerating.

• Analytics is now central to competitive advantage in business.

Lecture Notes – Framework for Business Intelligence (BI)

1. Introduction to BI Framework

• BI evolved from Decision Support Systems (DSS), Executive Information

Systems (EIS), and Management Information Systems (MIS).

• By 2006, the term Business Intelligence became standard for integrated tools,
methodologies, and systems supporting decision-making.

• BI enables:

o Interactive access to data (sometimes real-time)

o Manipulation of data

o Analysis for better decisions

• Process flow:
Data → Information → Decisions → Actions

2. Definitions of BI

• Umbrella term covering:

o Architectures

o Tools & analytical applications

o Databases

o Methodologies

• Major Objective: Turn raw data into actionable insights.

• Related terms:

o BPM – Business Performance Management

o Analytics, Data Science – Modern terms encompassing BI capabilities.

3. Brief History of BI

Era Development

1970s MIS → static 2D reporting (no analytics)

Early 1980s EIS → dynamic multidimensional reporting, forecasting, drill-down

Era Development

Mid-1990s Gartner coined "BI"; EIS evolved into BI

2005 onwards AI capabilities integrated into BI tools

Example: EIS dashboards of the 1980s evolved into today's interactive BI dashboards.

4. Architecture of BI

Four Major Components:

1. Data Warehouse (DW)

o Central repository for integrated, cleaned, and historical data.

2. Business Analytics

o Tools for data mining, OLAP, predictive analytics.

3. Business Performance Management (BPM)

o Monitoring KPIs, scorecards, dashboards.

4. User Interface
 o Dashboards, portals, visualizations.

Data Sources → ETL → Data Warehouse → Analytics Tools → Dashboards/Reports

5. Origins & Drivers of BI

• Drivers:

o Compliance requirements (e.g., Sarbanes-Oxley Act 2002)

o Faster business cycle times → need for quick, informed decisions

o Competitive advantage through data-driven operations

• Mantra: Right information → Right time → Right place

6. BI vs DSS

• DSS: Broader decision-making support, includes models and simulations.

• BI: Focus on data-driven reporting, analytics, collaboration.

• Some view DSS as a subset of BI, others as its precursor.

7. Transaction Processing vs Analytic Processing

• OLTP (Online Transaction Processing):

o Handles operational transactions (ATM, POS, ERP)

 o Optimized for fast updates, not complex analysis.

• OLAP (Online Analytical Processing):

o Works on DW data for multidimensional analysis.

o Supports ad hoc queries, drill-down, historical analysis.

8. Aligning BI with Business Strategy

• BI must be strategically aligned, not just a technical IT project.

• Gartner Planning Framework (2004):

1. Business & Organizational – Define strategic & operational goals.

2. Functionality – Choose analytics & reporting functions.

3. Infrastructure – Build scalable and integrated systems.

• BI Competency Center Functions:

o Link BI to strategy execution

o Share best practices

o Maintain standards of excellence

o Promote user–IT collaboration

9. Real-Time, On-Demand BI

• Need: Bridge gap between operational data and strategic goals.

• Approaches:

1. Fast DW updates – Near real-time ETL.

2. Business Activity Monitoring (BAM) – Direct event monitoring without

DW.

▪ Uses intelligent agents/web services to detect business events.

• Example: Automatic inventory reorder when stock drops below threshold.

10. Developing or Acquiring BI Systems

• Options:
 o Purchase ready-made BI tools

o Lease BI solutions

o Build custom BI systems

• Most BI solutions use vendor “shells” → customizable.

11. Justification & Cost–Benefit Analysis

• Include direct and intangible benefits.

• Benchmark against case studies ([Link] provides resources).

12. Security & Privacy in BI

• Protect strategic business data.

• Ensure employee and customer privacy compliance.

13. Integration of BI Systems

• BI must connect with:

o Databases

o ERP, CRM, SCM systems

o Web/e-commerce platforms

• All-in-one solutions simplify vendor management but may lose “best-of-breed”

flexibility.

Key Takeaways for Students

• BI is not just technology—it is a strategic initiative.

• Effective BI systems require alignment with business goals.

• Real-time BI is increasingly important due to fast-changing market conditions.

• Integration and data quality are critical for BI success.

Lecture Notes – Analytics Overview

1. Introduction to Analytics

• Definition:

o Analytics is the process of developing actionable decisions or

recommendations from insights derived from historical data.

o INFORMS definition: Combination of computer technology, management

science techniques, and statistics to solve real problems.

• Evolution of the Term:

o Replaced earlier terms like “Business Intelligence” (BI).

o Seen as a broader, modern umbrella term for computerized decision

support.

2. Levels of Analytics

According to INFORMS:
1. Descriptive Analytics – What happened? What is happening?

o Purpose: Understand trends, patterns, and causes in organizational data.

o Enablers: Business reporting, dashboards, scorecards, data

warehousing.

o Techniques: Reports, queries, alerts, trend analysis, visualization.

o Outcome: Informed understanding of current and past performance.

o Case Example (Silvaris): Used Tableau for real-time visualization and

decision-making.

2. Predictive Analytics – What will happen? Why will it happen?

o Purpose: Use historical data to forecast future events.

o Techniques: Data mining, statistical models, machine learning,

classification (logistic regression, decision trees, neural networks),
clustering, association mining.

o Applications: Customer churn prediction, recommendation systems

(e.g., Amazon product suggestions).

o Case Example (Purina – PANDA Platform):

▪ Built Azure-based platform to centralize scattered data.

▪ Developed predictive models to forecast cattle weight after 60

days with 2% error.

▪ Resulted in major efficiency, cost savings, and faster product

testing cycles.

3. Prescriptive Analytics – What should I do? Why should I do it?

o Purpose: Recommend optimal actions based on descriptive and

predictive insights.

o Techniques: Optimization, simulation, decision modeling, expert

systems.

o Outcome: Best possible decisions and actions for business problems.

o Case Example (Specialty Steel Bar Company):

▪ Developed real-time DSS using mixed-integer programming.

▪ Optimized available-to-promise (ATP) decisions for customer

orders.
3. Analytics in Practice – Case Study Lessons

Silvaris (Visualization)

• Challenge: Need for real-time data analysis in fluctuating markets.

• Solution: Tableau dashboards for live data interpretation.

• Lesson: Visualization is an effective entry point for real-time analytics.

Purina PANDA (Predictive)

• Challenges: Scattered, messy, varied data with no governance; lacked

centralized system.

• Solution: Built multi-layer architecture:

1. Data Acquisition Layer – Centralized raw data from multiple sources.

2. Transactional Layer – Converted data to uniform formats.

3. Analytical Layer – Applied logic, KPIs, statistical & ML models.

4. Visualization Layer – Dashboards & reports for decision-making.

• Impact: Saved $700K Capex, streamlined pipelines, increased efficiency,

accelerated product development.

Specialty Steel Bar (Prescriptive)

• Challenge: Optimize raw material allocation for customer orders.

• Solution: Real-time DSS with optimization models integrated into ERP.

• Lesson: Prescriptive analytics can automate and enhance high-stakes

operational decisions.

4. Analytics vs. Data Science vs. AI

• Overlap:

o Analytics includes descriptive, predictive, and prescriptive capabilities.

o Data Science often covers predictive & prescriptive, involving advanced

ML/AI.

• Skillset Differences:
 o Data Analyst: Descriptive/reporting analytics (Excel, SQL, basic
visualization).

o Data Scientist: Predictive & prescriptive analytics, statistical modeling,

programming (Python, R), ML/AI.

• Trend: AI increasingly used as the umbrella term for analytics and data science.

• Note: Boundaries between terms are blurry—usage often depends on academic

discipline or industry.

5. Key Takeaways

• Three Levels: Descriptive → Predictive → Prescriptive (increasing sophistication).

• Integration: Modern platforms unify all three types for better decision-making.

• Success Factors:

o Strong data governance.

o Collaboration between domain experts and analytics teams.

o Robust infrastructure for data collection, processing, and visualization.

• Outcome: Enhanced decision-making, efficiency, and competitive advantage.

Lecture Notes – 2.2 Introduction to Artificial Intelligence

1. Introduction

• Goal: Make computerized decision-making simpler, intuitive, and less

threatening.

• Example: A refrigerator that can evaluate contents and place online orders
automatically.

• AI is rapidly advancing and impacting many industries.

2. Economic Impact

• Projection (CIO Insight): By 2035, AI could create $5–$8.3 trillion in economic

value.
 • Key intelligent technologies:

o Internet of Things (IoT)

o Advanced robotics

o Self-driven vehicles

3. Definitions of AI

• AI focuses on:

1. Studying human thought processes.

2. Representing and duplicating these processes in machines.

• Merriam-Webster definition: Machine capability to imitate intelligent human

behavior.

• Subfield of computer science, with strong roots in logic.

• Early example: IBM Deep Blue beating Garry Kasparov in chess.

• Modern example: ChatGPT by OpenAI (natural language generation).

4. Major Characteristics of AI Machines

• Increasing intelligence in computing (e.g., Web 3.0 smarter than Web 2.0).

• Machine translation enabling cross-language communication and decision-

making.

5. Major Elements of AI

• Foundations: Philosophy, logic, mathematics, neurology, computer science,

sociology, psychology, linguistics, biology.

• Technologies & Applications:

o Machine learning

o Natural language processing (NLP)

o Computer vision

o Expert systems
 o Robotics

o IoT

o Deep learning

o Smart cities/factories/homes

o Augmented reality

o Intelligent agents

6. Applications of AI

• Smart systems can:

o Answer customer questions in natural language.

o Approve/reject purchase requests.

o Auto-generate orders.

o Support decision-making.

7. Case Study: Pitney Bowes

• Uses AI for automatic shipping fee calculation based on package dimensions,

weight, and packaging.

• Result: 25% improvement in accuracy → better pricing, customer satisfaction,

competitive advantage.

8. Major Goals of AI

• Create machines that:

o Reason

o Plan

o Solve problems

o Learn

o React to environmental changes

o Introduce creativity into processes

9. Drivers of AI

• Public interest in smart machines.

• Lower AI application costs vs. manual labor.

• Tech companies investing billions in AI.

• Pressure for higher productivity.

• Availability of quality data.

• Cheaper, more powerful computers.

• Cloud computing advances.

10. Benefits of AI

• Faster task completion.

• High consistency and accuracy.

• Continuous improvement.

• Predictive analysis & pattern recognition.

• Autonomous operation (no rest needed).

• Learning & performance improvement.

• Work in hazardous environments.

• Innovation support.

• No emotional bias.

• Fraud detection & security.

• Cost reduction & productivity increase.

• Improved decision-making & customer care.

11. Examples of AI Benefits

1. ISDA – Uses OCR & AI for contract digitization.

2. Recruitment – AI improves fairness and efficiency.

3. Management – Automates admin tasks; enables creative focus.

4. Accenture (India) – NLP & image recognition for visually impaired.

 5. Ford Motor Credit – Machine learning for better loan approvals.

6. IBM Watson – Predicts customer needs.

7. AI Startups – New businesses built around AI.

12. Limitations of AI Machines

• Lack human touch & emotional understanding.

• Limited environmental awareness outside tasks.

• Risk of over-reliance on AI.

• Potential misuse (destruction).

• Job displacement risk.

• Possibility of harmful autonomous behavior.

13. What AI Can & Cannot Do

• AI excels in simple input → simple output scenarios (e.g., loan approvals).

• Fully automated tasks are repetitive and data-driven.

• Barriers:

o Data unavailability/inaccuracy.

o Shortage of AI experts.

o High costs for certain applications.

14. Three Flavors of AI Decisions

1. Assisted Intelligence

o Weak AI in narrow domains.

o Examples: Monitoring systems, low-level personal assistants.

2. Autonomous AI

o Strong AI in specific domains, full decision-making authority.

o Examples: Robo-advisors, self-repairing robots, autonomous vehicles.

3. Augmented Intelligence
 o Extends human cognitive abilities for high performance.

o Current best option for practical organizational transformation.

15. Summary

• AI is transforming industries with its speed, accuracy, and adaptability.

• Offers enormous economic, operational, and social benefits.

• Must be managed carefully to reduce risks and ethical issues.

• Augmented intelligence currently provides the most practical and safe

approach.

Lecture Notes – 2.4 Major AI Technologies and Some Derivatives

1. Overview

• AI field is broad → Applications span from medicine to sports.

• Press (2017) identified Top 10 AI technologies and their life-cycle status.

• Major AI technologies relevant to business include:

o Intelligent agents

o Machine learning & deep learning

o Machine/computer vision

o Natural language processing (NLP)

o Knowledge systems

o Cognitive computing

o Augmented reality (AR)

2. Key AI Technologies

2.1 Intelligent Agents (IA)

• Definition: Small, autonomous software programs that sense environment

changes and act towards goals.

• Capabilities: May learn by expanding embedded knowledge.

• Initial Uses:

o Product search & recommendations

o Pricing determination

o Marketing planning

o Security enhancement

o Inventory management

• Benefits: Faster operations, cost reduction, fewer errors, improved customer

service.
 • Modern Applications: More sophisticated decision support, automation.

• Examples:

1. Virus detection software – scans incoming data, removes viruses, learns new
detection methods.

2. Allstate Business Insurance – helps agents provide complex rate quotes

quickly.

• Other Uses: Email filtering, news distribution, scheduling, automated data

gathering.

2.2 Machine Learning (ML)

• Definition: AI branch teaching computers to identify patterns and make

predictions using large datasets.

• Process:

o Data collection → Analysis → Model training → Prediction/Decision-

making.

• Applications: Fraud detection, predictive analytics, adaptive systems.

• Drivers of Growth:

o Increased computing power.

o Better algorithms.

o Deep neural networks.

o Big Data from IoT.

• Methods: Neural networks, case-based reasoning, etc.

Deep Learning (DL)

• Subset of ML that mimics human brain functions via artificial neural networks.

• Strengths: Handles complex, large-scale problems; continuous learning from

new data.

• Applications:

o Autonomous vehicles (road sign recognition).

o Robotics.
 o Speech & voice processing.

o Smart homes, cities.

o Stock market predictions.

• Example: Cargill Corp uses DL to improve shrimp farming efficiency.

2.3 Machine Vision & Computer Vision

• Machine Vision: Imaging-based inspection & analysis for industrial uses (robot
guidance, process control, quality inspection).

• Computer Vision: Enables computers to interpret and understand

images/videos; applied in safety, security, healthcare, entertainment.

• Differences:

o Machine vision: engineering focus.

o Computer vision: computer science focus.

• Applications:

o Scene recognition.

o Facial recognition (e.g., Chinese police identifying suspects in large

crowds).

o Illegal logging detection.

o Wildlife poaching prevention.

• Video Analytics: Uses vision techniques to detect patterns/events in video

streams.

2.4 Natural Language Processing (NLP)

• Definition: Enables human-computer interaction in natural languages (text or

speech).

• Two Subfields:

1. Natural Language Understanding (NLU) – Computer comprehension of

human language.

2. Natural Language Generation (NLG) – Computers produce human-

understandable language.
 • Related Technologies:

o Speech Understanding: Voice recognition & interpretation (e.g., call

center automation).

o Machine Translation: Automated language translation (Google Translate,

Babel Fish), including camera-based instant translation.

2.5 Knowledge Systems

• Knowledge Acquisition:

o Identify required knowledge.

o Determine sources (documents, experts, data).

o Extract & structure knowledge.

o Validate and verify.

• Knowledge Representation:

o Organize for reasoning and interaction.

o Examples: Q&A format, rules, semantic networks.

• Reasoning from Knowledge:

o Core function for problem-solving and recommendations.

o Differentiates types of intelligent systems.

2.6 Cognitive Computing

• Definition: Combines cognitive science & AI to simulate human thought

processes for decision support.

• Features:

o Self-learning algorithms.

o Pattern recognition.

o NLP, vision, reasoning.

• Example: IBM Watson supporting complex decision-making.

2.7 Augmented Reality (AR)

• Definition: Real-time integration of digital information into the user’s physical

environment.

• Technologies Used: Machine vision, scene recognition, gesture recognition,

sensors.

• Applications:

o Real estate (instant property data).

 o Education.

o Gaming.

o Retail experience enhancement.

• Platforms: Apple’s ARKit, Android ARCore.

3. Summary Table – AI Technologies & Business Relevance

Technology Purpose Example Application

Intelligent Agents Autonomous task execution Call center support (Allstate)

Machine Learning Pattern detection & prediction Credit card fraud detection

Deep Learning Complex, continuous learning Autonomous vehicles

Machine Vision Industrial inspection Quality control

Computer Vision Image/video understanding Facial recognition

NLP Language interaction Chatbots, translation

Knowledge Systems Reasoning from knowledge Expert decision support

Cognitive Computing Human-like decision-making IBM Watson

Augmented Reality Digital-physical integration Interactive retail displays

This format works well for lectures because:

• It’s broken into logical sections.

• Uses clear definitions + examples.

• Includes applications for relevance.

• Summarized in a table for quick recall.

Lecture Notes: AI Applications in Business

Introduction

• AI is transforming traditional business areas such as:

o Accounting

o Finance & Banking

o Insurance

o Human Resource Management (HRM)

o Marketing, Advertising, & CRM

o Production, Operations & Logistics

1. AI in Accounting

Overview

• Major users: Large tax & accounting firms (EY, Deloitte, PwC).

• Tools: NLP, Machine Learning, Robotic Process Automation (RPA), Text Mining.

• Used for: Tax preparation, auditing, consulting, contract reviews.

Big Firms’ Strategies

• EY: Focus on ROI; e.g., reviewing leasing documents for compliance.

• PwC: Quick, small projects (4 weeks); demonstrates value (70–80 projects

annually).

• Deloitte: Innovation-driven; e.g., NLP for contract review, M&A data analysis,
procurement synergies.

Small Firms

• Example: Crowe Horwath (Chicago) uses AI for healthcare billing and claims.

• AI assists in risk analysis, real estate contract analysis, and reimbursements.

Impact

• Automates routine accounting tasks.

• Enhances accuracy and reduces processing time.

• Job Impact: Some routine jobs may decline, but accountants will manage AI
systems & drive innovation.
2. AI in Financial Services

Key AI Activities (Singh, 2017)

• Extreme personalization (chatbots, robo-advisors).

• Shifts in customer behavior (online & branch).

• Trust in digital identity.

• Payment revolution.

• Peer-to-peer lending & sharing economy.

• 24/7, global financial services.

Banking Applications

• Employee surveillance: Detect fraud (e.g., Wells Fargo).

• Tax preparation: H&R Block + IBM Watson → ensures accurate tax returns.

• Customer queries: Rainbird AI bots assist in answering millions of queries.

• Voice assistants: Capital One with Alexa → pay bills, check balances.

• Fraud detection: Bank Danamon uses ML for AML (Anti–Money Laundering).

• Virtual assistants: HSBC’s Olivia, Santander’s Nina, RBS’s Luvo.

• Mortgage advising: Accenture’s Collette.

• Nao robot: Detects nationality via facial recognition, switches language.

Insurance Applications

• AI supports policy issuance & claims management.

• NLP & text recognition analyze incoming claims.

• AI reduces claim processing time by up to 90%.

• AI used in underwriting, fraud detection, risk analysis.

• Benefits: Faster, accurate, efficient, and scalable globally.

3. AI in Human Resource Management (HRM)

Reasons for AI in HRM (Savar, 2017)

1. Reduce human bias.

2. Improve efficiency and productivity.

3. Enhance employee relationships.

Areas of Application (Wislow, 2017)

• Recruitment: LinkedIn AI suggests matches, reducing unconscious bias.

• Training & Learning: Chatbots answer queries, AI personalizes online courses,

progress testing.

• Performance Analysis: Breaks tasks into components, measures employee &

team output.

• Retention & Attrition: ML predicts employee turnover and suggests retention

strategies.

• Onboarding: AI creates customized employee onboarding paths → increases

retention.

• Chatbots in HR:

o Mya → recruitment assistant.

o Job Bot → supports hiring hourly workers (e.g., via Craigslist).

4. AI in Marketing, Advertising, and CRM

Customer Experience & CRM

• Conversational bots (e.g., Alexa) enhance interaction.

• Example: Salesforce Einstein

o Vision recognition, predictive analytics, sales dashboards.

o Provides insights on leads, customer sentiment, competitors.

o Prioritizes sales leads, boosting sales productivity.

Other Applications

• Mimicking salespeople: Robots in stores guide customers.

• Lead generation: Analyzes customer data, predicts sales opportunities.

 • Customer loyalty: Personalization (e.g., IBM Watson learns from tweets).

• Sales pipeline improvement (Narayan, 2018):

1. Identify target customers.

2. Send personalized ads/videos.

3. Deliver leads to sales teams.

5. AI in Manufacturing & Operations

Manufacturing

• AI + Robotics = Mass customization & make-to-order products.

• Intelligent robots replace repetitive work.

• Defect detection: Sensors + AI instantly remove faulty items.

• Predictive maintenance: Prevents equipment failures → saves cost.

Logistics & Supply Chain

• Example: Amazon → 50,000+ robots in warehouses.

• Example: DHL Supply Chain

o Uses AI, IoT, ML for supply chain optimization.

o Applications: Inventory management, warehouse automation, RFID & QR

tracking, predictive analytics.

o Enables faster procurement, planning, scheduling, and resource

optimization.

Conclusion

• AI is not just automating routine tasks but also transforming business models.

• Across industries: AI improves efficiency, personalization, decision-making,

and customer experience.

• Future challenge: Balancing automation with workforce adaptation and ethics.

Lecture Notes: Conversational AI – Chatbots

1. Introduction to Chatbots

• Definition:

o Chatbot = "chat robot" (also called bot or robo).

o Computerized service that enables human–computer conversations.

o Communication modes: text, voice, images.

o Uses NLP (Natural Language Processing) to understand unstructured

human dialog.

• Purpose:

o Information retrieval, personal assistance, customer service,

communication, collaboration.

• Trend:

o Widespread adoption in recent years → “Chatbot Revolution.”

o Common in business, healthcare, marketing, finance, government

services, manufacturing.

o More personal and engaging than traditional systems.

2. Evolution of Chatbots

• Early Stage:

o Originated as simple expert systems decades ago.

o Example: ELIZA (1960s) → worked in Q&A mode with FAQs.

o Limitations: irrelevant or misunderstood answers, narrow domains.

• Later Development:

o 2000s: Improved NLP, better knowledge bases (cloud-based, cost-

effective).

o 2010s: Term “chatbots” became popular; evolved into Virtual Personal

Assistants (e.g., Amazon Alexa, Google Assistant).
3. Types of Bots

1. Regular Bots

o Perform simple, repetitive tasks (e.g., show bank debits, buy stocks, shop
online).

2. Chatbots

o More capable, simulate natural conversations with humans.

3. Intelligent Bots

o Equipped with learning capability.

o Improve knowledge base over time (e.g., Alexa, robo-advisors).

4. Drivers of Chatbots

• Technology Drivers:

o Advances in AI, NLP, ML, DL, and cloud-based knowledge systems.

o Tools for rapid and low-cost chatbot development.

• Business Drivers:

o 24/7 customer service.

o Cost reduction & scalability without hiring many employees.

o Enhanced customer engagement and marketing.

o Integration with messaging apps (e.g., FB Messenger, WeChat).

5. Components of a Chatbot

• User (Client)

• AI Machine (computer/avatar/robot)

• Knowledge Base (local or cloud-based)

• Human–Computer Interface (text/voice interaction)

• NLP Engine (understand queries)

• Optional Features: Gesture and voice tone recognition.

6. Chatbot Conversation Process (Fig. 2.6)

1. User asks a question (text/voice).

2. NLP converts query into machine-readable form.

3. Query sent to cloud services (knowledge base, analytics, business logic).

4. Response generated → passed through Natural Language Generation (NLG).

5. Answer delivered back to user in preferred mode (text, voice, etc.).

7. Benefits of Chatbots

• Cost reduction.

• Multilingual communication (via machine translation).

• Improved knowledge quality & response accuracy.

• Superior customer service and marketing insights.

• Applications: text/image recognition, shopping, decision support.

• Healthcare uses: robot receptionists, dementia therapy, remote diagnostics.

8. Representative Chatbots Worldwide

• RoboCoke (Coca-Cola, Hungary) → Music & party recommendations.

• Kip (Slack) → Shopping assistant bot.

• Walnut → Identifies and teaches relevant skills.

• Taxi Bot → Compares Uber, Lyft, Grab, etc., with promo codes.

• ShopiiBot → Finds products from images or descriptions.

• Tour/Trip Bots → Suggests restaurants, events, attractions.

• BO.T (Bolivia) → Answers cultural, geographical, social queries in Spanish.

• Hazie → Career and job search assistant.

• Visabot (Green Card) → Assists with US Green Card applications.

9. Key Takeaways

• Chatbots = conversational AI systems driven by NLP and knowledge bases.

• Evolved from simple FAQ systems → intelligent assistants.

• Offer businesses cost savings, scalability, and enhanced user engagement.

• Widely used in healthcare, e-commerce, marketing, travel, government, and

finance.

• Represent a revolution in digital human–machine communication.