Module 1

1. Role of Analytics in Decision Making

Organizations use analytics to:

- Develop reports: Understand current operations.

- Predict trends: Forecast future outcomes.
- Make decisions: Optimize actions for desired results.

Modern organizations handle large-scale data for informed decision-making.

Computerized systems are applied in:

- Payroll and bookkeeping

- Automated factories management
- Evaluation of mergers and acquisitions

2. Importance of Decision Making

Central to organizational success; wrong decisions can be costly.

Increasing complexity due to internal and external factors.

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

1. Big-bet, high-risk decisions

2. Cross-cutting decisions – repetitive, high-risk, require group work
3. Ad hoc decisions – episodic, situational
4. Delegated decisions – assigned to individuals or small teams

3. Evolution of Decision Making

Traditionally: Relied on intuition, creativity, and experience.

Modern approach: Focus on methodical, analytical decision-making.

Computers support managers by:

- Data collection and storage

- Simulation and forecasting
- Automating routine decisions

1
4. Technologies Supporting Decision Making
1. Group Communication & Collaboration
- Remote collaboration via tools and smartphones.
- Essential for supply chain responsiveness.
2. Improved Data Management
- Fast access to diverse data (text, graphics, sound, video).
3. Big Data & Data Warehousing
- Tools: Hadoop, Spark, cloud platforms handle massive datasets.
4. Analytical Support
- Evaluate alternatives, perform risk analysis, gather expert input.

5. Overcoming Cognitive Limits

- Systems help humans process and recall large data efficiently.

6. Knowledge Management

- Extract value from structured/unstructured organizational data.

7. Anywhere, Anytime Support

- Mobile/cloud technologies enable on-the-go decision-making.

8. Innovation & AI

- AI integrated with analytics enhances decision-making at every step.

5. Simon’s Decision-Making Process (1977)

Four Phases:

1. Intelligence Phase
- Identify and define problems/opportunities.
- Determine problem ownership.
- Collect data and distinguish real problems vs. symptoms.
- Example: Elevator waiting time – real issue was perceived waiting time, solved with mirrors.
2. Design Phase
- Develop models representing the problem.
- Test feasibility and predict outcomes.
- Simplify reality using assumptions to balance cost and accuracy.
3. Choice Phase
- Select the best solution from alternatives.
- Perform sensitivity analysis, what-if analysis, and goal-seeking.
4. Implementation Phase
- Execute the chosen solution and manage change.
- Monitor results and collect feedback for continuous improvement.
- Incorporate project management principles and manage user expectations.

2
Feedback loops: Each phase may return to a previous phase to refine decisions.

6. Data Issues in Decision Making

Challenges include:

- Data unavailability or high cost

- Inaccuracy or imprecision
- Information overload
- Estimations and assumptions may be subjective
- Future data may not mirror historical trends
- Qualitative (soft) data may be critical

7. Problem Classification and Ownership

- Structured problems: Well-defined, standard solutions.
- Unstructured problems: Require creativity and analysis.

3
 - Semi-structured problems: Mix of structured and unstructured subproblems.
- Problem decomposition: Break complex problems into simpler subproblems.
- Problem ownership: Responsibility assigned to a person or team for resolution.

Key Takeaways
- Decision-making is systematic and data-driven, not just intuition-based.
- Technology, analytics, and AI are essential enablers.
- Correctly identifying the real problem is crucial for effective solutions.
- Continuous feedback and data analysis improve future decisions.

Evolution of Decision Support, Business Intelligence, Analytics, and

AI

1. Introduction to Analytics Evolution

 Since the 1970s, analytics terminology has evolved significantly.
 1970s: Focus on structured, periodic reports (e.g., daily, weekly, monthly,
quarterly).
 These reports, termed Management Information Systems (MIS), helped managers
understand past performance but lacked flexibility.

2. Decision Support Systems (DSS)

 Early 1970s: Scott-Morton defined DSS as interactive computer-based systems
helping decision makers solve unstructured problems.
 Keen & Scott-Morton (1978): DSS couples human intellect with computer
capabilities for semi structured problems.
 DSS helped utilize data and models for better decisions.

3. Operations Research and Expert Systems

4
 Early analytics relied on domain experts, using interviews/surveys to build models
(Operations Research, OR).
 Complex problems handled via heuristics, simulations (prescriptive analytics).
 Late 1970s–1980s: Expert Systems (ES) captured expert knowledge via if–then–
else rules, enabling intelligent DSS.

4. Enterprise Resource Planning (ERP) and Relational Databases (1980s)

 Integrated previously disjointed information systems (finance, HR, marketing).
 RDBMS replaced sequential/nonstandard data, improving storage, relationships,
and reducing duplication.
 ERP provided a single version of truth across organizations.
 Enabled on-demand, specialized reporting for decision makers.

5. Executive Information Systems (1990s)

 Developed for executives: visual dashboards and scorecards to track KPIs.
 Data warehouses (DW) introduced to support reporting without affecting
transactional systems (ERP).

6. Business Intelligence (2000s)

 DW-driven DSS evolved into BI systems.
 Real-time/right-time data warehousing addressed data latency.
 Data mining and text mining extracted insights from growing DWs.
 Service-oriented architectures enabled small/medium enterprises to access
analytics as a service.

7. Big Data Era (2010s)

 New data sources: RFID, smart devices, wearable health monitors, social media.
 Big Data: challenges in volume, variety, and velocity.
 Advances: Hadoop, MapReduce, NoSQL, parallel processing.
 Enabled streaming analytics, IoT, and real-time insights.

8. AI and Modern Analytics

 AI integrated with BI to improve decision-making.
 Deep learning enabled new data analyses (images, voice, gesture interfaces).
 Cloud-based storage and high-speed processing enabled mass data handling.
 Tools like ChatGPT showcase AI-driven analytics in real-time interaction.

9. Impact of COVID-19
 Highlighted need for data-driven decisions and predictive modeling.
 Economic slowdown caused temporary reduction in analytics investments.

5
 Post-pandemic: renewed focus on analytics for efficiency, revenue growth, fraud
detection, customer engagement.

Business Intelligence (BI) – Concepts and Implementation

1. Evolution of Decision Support and BI
 Decision support concepts were implemented under various names by different
vendors.
 EISs (Executive Information Systems) provided user-friendly reports for faster
decision-making.
 By 2006, these evolved into Business Intelligence (BI), adding:
o Visualizations
o Alerts
o Performance measurement capabilities

2. Definition of BI
 BI is an umbrella term for:
o Architectures, tools, databases, analytical applications, and methodologies
 Goal: Enable managers and analysts to access, manipulate, and analyze data for
better decisions.
BI transforms data → information → decisions → actions.

6
BI Architecture
Four main components:
1. Data Warehouse (DW): Centralized storage of integrated data from multiple
sources.
2. Business Analytics: Tools for mining, manipulating, and analyzing DW data.
3. Business Performance Management (BPM): Monitoring and analyzing
performance.

7
4. User Interface: Dashboards, portals, browsers for accessing insights.

Drivers and Importance of BI

 Captures and harnesses data to support decision-making and improve business
operations.
 Legislative compliance (e.g., Sarbanes-Oxley Act) requires accurate reporting.
 Fast and informed decision-making is crucial due to compressed business cycles.
 Companies often establish a BI Competency Center to:
o Align BI with strategy
o Promote collaboration between IT and business users
o Share best practices
o Standardize BI processes
o Educate stakeholders on BI benefits

Real-Time and On-Demand BI

 Users demand instant access to data for quick decisions.
 Traditional BI: Static, batch-processed data in DW.
 Real-time BI approaches:
1. Near-real-time DW updates (faster ETL processes)
2. Business Activity Management (BAM):
 Event- and process-based monitoring
 Proactive alerts for operational changes
 Applications: SCM auto-orders, CRM triggers alerts for large orders

Developing or Acquiring BI Systems

 Options: Purchase, lease, or build BI tools.
 Decisions depend on:

8
 o Cost–benefit analysis
o Justification of direct and intangible benefits
 Resources: TDWI (tdwi.org) for BI product information and case studies.

Key Considerations
1. Security & Privacy: Protect strategic, employee, and customer data.
2. System Integration: BI must integrate with ERP, CRM, legacy systems, e-
commerce, and partner systems.
3. Alignment with Strategy: BI should improve business processes and transform
decision-making.
4. Planning Framework (Gartner, 2004):
o Business objectives
o Organizational readiness
o Functionality
o Infrastructure
Real-Time Capability: Supports agile, on-demand decision-making with alerts,
notifications, and automated actions.

Transaction Processing vs. Analytic Processing

 OLTP (Online Transaction Processing):
o Handles routine operations efficiently (e.g., sales, deposits, inventory)
o Poor for ad-hoc reporting
 OLAP (Online Analytic Processing):
o Uses DW data for analysis and decision-making
o Supports tactical and operational decisions
 Separation of OLTP and OLAP enables effective BI.

BI vs. DSS
 DSS: Supports semi-structured decisions.
 BI: Includes DSS as part of a broader system; focuses on reporting, collaboration,
and communication.
 BI represents an evolution of DSS for modern analytics needs.

Summary
 BI integrates data, analytics, visualization, and decision-making tools.
 Real-time, integrated, and strategically aligned BI provides a competitive
advantage.
 Planning, governance, integration, and security are key to successful BI
implementation.

9
 1. Introduction
 The term analytics has largely replaced older decision support technologies,
including BI.
 Analytics involves developing actionable decisions or recommendations from
historical data.
 INFORMS defines analytics as the combination of computer technology,
management science, and statistics to solve real problems.
 SAS Institute outlines eight levels of analytics, starting from standardized reports
to predictive and prescriptive insights.

Levels of Analytics
Analytics is typically classified into three levels:
Descriptive Analytics
 Answers: “What happened?” and “What is happening?”
 Involves consolidation of data sources into a data warehouse (DW).
 Tools: Business reporting, dashboards, scorecards, visualization.
 Purpose: Understand trends, causes, and current performance.
 Outcome: Provides insight into business operations.
Predictive Analytics
 Answers: “What will happen?” and “Why will it happen?”
 Uses statistical and data mining techniques.
 Purpose: Predict future events like customer behavior, churn, or sales trends.
 Techniques:
o Classification: Logistic regression, decision trees, neural networks
o Clustering: Segment customers for targeted actions
o Association mining: Identify relationships between products or behaviors
 Outcome: Provides accurate projections of future events.
Prescriptive Analytics
 Answers: “What should I do?” and “Why should I do it?”
 Uses optimization, simulation, decision modeling, and expert systems.
 Purpose: Recommend best possible decisions and actions.
 Outcome: Enables well-informed business decisions and actionable strategies.

Key Features Across Analytics Levels

 Descriptive: Reporting and visualization to understand data.
 Predictive: Data mining and modeling to forecast outcomes.
 Prescriptive: Decision optimization to guide actions.
 These levels are interconnected: descriptive insights feed predictive models,
which inform prescriptive decisions.
10
Enablers of Analytics
 Data Infrastructure: Data warehouses for consolidated, accessible data.
 Visualization Tools: Enable easier interpretation and actionable insights.
 Advanced Algorithms: Statistical, machine learning, and AI methods.
Outcomes
 Descriptive: Understand current performance and trends.
 Predictive: Forecast future trends and behaviors.
 Prescriptive: Determine optimal decisions to achieve desired outcomes.

Artificial Intelligence (AI) is a vast field with applications across numerous disciplines,
including medicine, sports, business, and more. AI technologies are evolving rapidly,
and several have become critical in business applications. Some of the major AI
technologies include autonomous vehicles, image generation, robotics, augmented
reality, machine vision, knowledge systems, video analysis, cognitive computing,
chatbots, natural language processing, machine learning, intelligent agents, deep
learning, speech and voice understanding, machine translation, and neural networks.

 Intelligent Agents (IA) are autonomous software programs that observe and act on
changes in their environment by performing tasks independently. IAs are designed to
achieve specific goals by monitoring and responding to environmental changes. They

11
 can also learn by expanding the knowledge embedded within them. Intelligent agents
help overcome internet information overload and make computers more effective as
decision-support tools. Initially, IAs were used for routine tasks such as product
searches, recommendations, pricing, marketing planning, security, auction
management, payment facilitation, and inventory management, primarily to increase
speed, reduce costs and errors, and improve customer service.
 Examples of intelligent agents include virus detection programs that scan and remove
malware while learning new detection methods, and Allstate Business Insurance using
agents to assist human agents in complex rate-quoting processes. Other applications
include e-mail servers, news filtering, appointment management, and automated
information gathering.
 Machine Learning (ML) is a subfield of AI where systems learn from data rather than
having explicit programming for every task. ML algorithms help computers recognize
patterns, make predictions, and support decision-making. Machine learning systems
monitor their environment, adjust behavior based on changes, and can reconfigure
programs as conditions evolve. It involves training algorithms on large datasets to
identify hidden patterns or relationships. Applications of ML include detecting credit
card fraud, performance prediction, and predictive analytics. Advances in computing
power, algorithms, deep neural networks, and access to big data (e.g., IoT sources) have
fueled the rapid growth of machine learning.
 Deep Learning (DL) is a specialized subset of machine learning designed to mimic
human brain functioning using artificial neural networks. Deep learning can handle
complex applications that traditional ML cannot, such as autonomous vehicle
navigation, robotics, machine vision, scene recognition, and speech/voice processing.
DL systems continuously learn from incoming data, enabling self-directed adaptation.
Examples include DL-based analytics in agriculture for improving production
efficiency and stock market predictions.
 Machine and Computer Vision involve systems that allow computers to process and
interpret visual information. Machine vision focuses on automated imaging-based
inspections and analysis for industrial applications like robot guidance and quality
control, while computer vision is an interdisciplinary field that seeks high-level
understanding from images or videos. Both technologies automate tasks, reduce labor
costs, and enhance decision-making. Applications include scene recognition, illegal
logging detection, facial recognition for security, and monitoring wildlife poaching.
 Video Analytics applies computer vision techniques to videos for pattern recognition,
fraud detection, and event identification. This can also be used to predict human
interactions or advertising effectiveness.
 Natural Language Processing (NLP) enables computers to understand, interpret, and
generate human language, allowing communication through text or speech.

12
 NLP includes two main subfields: natural language understanding (enabling
comprehension of queries or instructions) and natural language generation (producing
human-readable text).
 Speech and Voice Understanding allows computers to recognize and interpret spoken
language, commonly applied in automated call centers and virtual assistants.
 Machine Translation involves translating words and sentences from one language to
another using computer programs. Modern applications, like Google Translate,
combine translation with vision (e.g., translating text via smartphone camera) and
speech recognition.
 Knowledge Acquisition and Representation are essential for intelligent systems to
function. Knowledge acquisition involves identifying, extracting, and structuring
necessary information, often requiring expert cooperation. Knowledge is then
organized and stored in a knowledge repository for reasoning, Q&A, problem-
solving, or decision-making.

 Reasoning from Knowledge is the core of intelligent systems, processing user

requests and providing solutions, recommendations, or insights.
 Cognitive Computing combines knowledge from cognitive science and computer
science to simulate human thought processes, enabling computers to support complex
decision-making and problem-solving. These systems use self-learning algorithms,
pattern recognition, NLP, and machine vision, exemplified by IBM Watson.

13
 Augmented Reality (AR) integrates digital information into the real world in real-
time, enhancing interactive experiences. AR combines machine vision, scene
recognition, and sensor data (vision, sound, temperature) to provide actionable insights.
Applications include augmented information on real-world objects, improving
learning, work, and purchasing decisions. ARKit on iPhones is an example of this
technology in use today.

AI technologies are being widely adopted in various business domains, including accounting,
finance, human resources, marketing, production, and logistics. They help companies automate
tasks, improve efficiency, and deliver better experiences to customers and employees.

AI in Accounting
AI is transforming accounting by automating repetitive tasks and enhancing decision-making:
 Large Accounting Firms: Companies like EY, Deloitte, and PwC use AI to support or
replace human activities in tax preparation, auditing, and consulting. Tools include
Natural Language Processing (NLP), robotic process automation, text mining, and
machine learning. These tools reduce processing times and improve accuracy in document
review, compliance, and financial analysis.
 Small Firms: Smaller accounting companies also benefit from AI. For example, Crowe
Horwath uses AI to handle complex billing and claims processing in healthcare. AI helps
firms analyze contracts, assess risk, and streamline processes.
 Impact on Accountants: Routine tasks are automated, requiring accountants to manage
AI systems, focus on higher-level decision-making, and drive innovation in AI
applications.

AI in Financial Services
Financial services use AI to enhance personalization, trust, efficiency, and customer experience.
Key applications include:
 Banking:
o Employee surveillance to prevent fraud and misconduct.
o Automated tax preparation and advice (e.g., H&R Block using IBM Watson).
o Real-time customer query handling with chatbots and virtual assistants (e.g.,
Capital One with Alexa, Santander Bank with Nina).
o Fraud detection and anti-money laundering using machine learning (e.g., Bank
Danamon).
o Facial recognition and behavior analysis to customize customer interactions.
 Insurance:
o Claims processing and policy issuance are streamlined using AI.
o AI can analyze claims, categorize them, and route them to the appropriate
adjusters.
o Technologies like NLP and text recognition enhance data collection, underwriting,
and fraud detection.

14
 o AI reduces processing time, improves accuracy, and supports global multi-office
configurations.
 AI in finance increases personalization, enabling services to be tailored to individual customer
needs while improving operational efficiency and reducing costs.

AI in Human Resource Management (HRM)

AI improves recruiting, training, performance evaluation, and employee retention:
 Recruitment: AI algorithms match candidates to job openings, reducing human bias and
improving efficiency in talent acquisition. Platforms like LinkedIn use AI for candidate
recommendations.
 Training and Learning: AI facilitates personalized learning, online courses, chatbots for
queries, and progress tracking to support employee development.
 Performance Analysis: AI breaks work into measurable components, evaluates
performance against objectives, and tracks changes over time.
 Retention and Attrition Detection: Machine learning identifies patterns in employee
departures and helps predict retention strategies.
 Onboarding: AI designs customized onboarding plans for new hires, improving
engagement and long-term retention.
 HR Chatbots: Assist employees with queries anytime, examples include Mya (recruiting)
and Job Bot (hourly worker recruitment).

AI in Marketing, Advertising, and CRM

AI enhances customer experiences, supports sales, and personalizes interactions:
 Customer Experience and CRM: Conversational bots provide information, guidance,
and advice. Salesforce’s Einstein delivers AI-driven insights, manages sales dashboards,
prioritizes leads, and predicts customer behavior.
 Sales and Lead Generation: AI analyzes customer data, generates leads, predicts
conversions, and helps sales teams focus on high-value prospects.
 Personalization and Loyalty: AI tracks customer behavior and tailors experiences, offers,
and recommendations. Tools like IBM Watson analyze social media and purchase behavior
to improve loyalty.
 In-Store Guidance: AI-powered robots and assistants help customers in physical stores,
providing expertise and support.

AI in Manufacturing and Operations

AI improves efficiency, reduces labor costs, and enables real-time decision-making in production:
 Automation and Robotics: Intelligent robots perform complex tasks, enabling mass
customization and make-to-order production.
 Quality Control and Predictive Maintenance: Sensors detect defective products or
equipment malfunctions. AI algorithms take immediate corrective actions or predict
failures to minimize downtime and costs.
 Agility and Customer Experience: AI allows manufacturers to respond quickly to
changing demands, improve product quality, and optimize operational efficiency.

15
AI in Logistics and Transportation
AI and robotics optimize supply chains, warehousing, and delivery operations:
 Warehouse Management: Companies like Amazon use tens of thousands of robots for
sorting and transporting goods within distribution centers.
 Predictive Analytics: AI forecasts demand, plans procurement, and schedules resources
more accurately.
 Supply Chain Visibility: RFID and QR code tagging allow real-time tracking of items
along the supply chain.
 Innovative Business Models: Companies like DHL integrate AI, IoT, and machine
learning to optimize logistics, inventory management, and transportation for themselves
and their clients.

AI Benefits in Business
 Efficiency: Automates repetitive, high-volume tasks and reduces processing times.
 Accuracy: Minimizes human errors in tasks like auditing, claims processing, and quality
control.
 Personalization: Delivers tailored customer experiences in finance, marketing, and CRM.
 Decision Support: Enhances human decision-making using predictive analytics and
cognitive computing.
 Cost Reduction: Reduces labor costs, operational waste, and downtime through
automation and predictive maintenance.

Chatbots are computerized systems that simulate conversation with humans. They have
become increasingly popular in recent years due to advancements in artificial
intelligence (AI), particularly in natural language processing (NLP), machine
learning, deep learning, and knowledge systems. Chatbots can interact via text, voice,
or images and are used in business, healthcare, government, and personal applications.

What Is a Chatbot?
 Short for chat robot, a chatbot (or “bot”) enables easy communication between humans
and computers.
 Chatbots often use NLP to understand unstructured human language.
 They can stand alone or integrate with other knowledge systems.
 Interactions can occur over messaging platforms, web pages, mobile apps, and social media
(e.g., Facebook Messenger, WeChat, Twitter).

Evolution of Chatbots
1. Early Bots: Simple Q&A expert systems (e.g., ELIZA) that matched questions to pre-
stored answers. Limitations included narrow domain knowledge and poor understanding
of language.
2. Modern Bots: Conversational AI with the ability to learn from interactions, integrate
cloud-based knowledge, and provide personalized responses.

16
 3. Virtual Personal Assistants: Examples include Amazon Alexa and Google Assistant,
which combine chatbots with intelligent agents and voice recognition.

Types of Bots
1. Regular Bots: Perform simple, repetitive tasks (e.g., checking bank balances,
buying/selling online).
2. Chatbots: Engage in human-like conversations and assist in tasks such as customer service
and recommendations.
3. Intelligent Bots: Continuously learn from interactions and user preferences (e.g., Alexa,
robo-advisors).

Drivers of Chatbots
 Availability of tools to build bots quickly and cost-effectively.
 Improved AI capabilities, especially NLP and voice technologies.
 Growing demand for 24/7 customer service and marketing automation.
 Reduced need for large customer support teams.
 Integration with popular messaging platforms and mobile devices.

Components of Chatbots
1. Person (Client): The human interacting with the bot.
2. AI Machine/Avatar: The chatbot that responds.
3. Knowledge Base: Stores information; can be local or cloud-based.
4. Human–Computer Interface: Supports text, voice, or gesture communication.
5. NLP: Allows the bot to understand natural language.
6. Advanced Features: Gesture recognition, voice variations, image input.

How Chatbots Work

1. The user asks a question via text, voice, or images.
2. NLP translates the query into machine-understandable language.
3. The bot queries its knowledge base and analytics in the cloud.
4. The response is generated through Natural Language Generation (NLG).
5. The answer is delivered to the user in their preferred communication mode.

Benefits of Chatbots
 Cost reduction by automating tasks.
 Improved customer service and marketing efficiency.
 Multi-language support through machine translation.
 Enhanced knowledge capture and sharing.
 Decision support, shopping assistance, and market research.
 Personalized interaction and therapy applications (e.g., Nao robot for elderly or dementia
care).

Representative Chatbots Worldwide

 RoboCoke: Music and party recommendations (Coca-Cola, Hungary).

17
  Kip: Shopping assistant on Slack.
 Walnut: Skill discovery and learning guidance.
 Taxi Bot: Compares ride-sharing options and promo codes.
 ShopiiBot: Finds products via images or search queries.
 BO.T: Bolivian cultural and geographical information in Spanish.
 Hazie: Career guidance and job placement assistance.
 Green Card (Visabot): Assists users with filing US Green Card requests.

Why Chatbots Are AI Machines

 They use machine learning to improve responses over time.
 They rely on knowledge systems and analytics for decision-making.
 They can understand, process, and generate human language.
 They provide intelligent services across multiple industries and tasks.

18