Key Concepts in IT Architecture and AI Technologies

Technical Debt

The implied long-term costs of choosing quick, short-term IT solutions (e.g., patching old code) instead of investing in sustainable, scalable systems. Over time, this 'debt' slows progress and increases maintenance costs.

IT Architecture

A structured framework that defines how an organization's IT systems (hardware, software, networks) are integrated to support business processes and goals.

Information Systems Planning Process

Steps: Align IT with business goals. Assess current IT infrastructure. Define architecture (e.g., cloud vs. on-premise). Prioritize projects (e.g., ROI analysis). Implement and monitor.

Technical Aspects of IT Architecture (Application Software)

Examples: Enterprise systems (ERP, CRM). Middleware (connects applications). Cloud-based apps (SaaS like Salesforce).

IS Strategic Plan

A long-term roadmap outlining how IT investments will align with and enable business objectives (e.g., digital transformation).

Application Portfolio

A curated inventory of all software applications used by an organization, categorized by function (e.g., HR, finance) and value to prioritize resources.

Fixed vs. Variable Costs

Fixed: Unchanged with usage (e.g., executive salaries, server leases). Variable: Scale with usage (e.g., cloud storage fees, per-user SaaS licenses).

Return on Investment (ROI)

A financial metric measuring profitability of an investment. Formula: (Net Profit / Cost of Investment) × 100%. Measures: Efficiency of capital allocation (e.g., investing in new software).

Software as a Service (SaaS)

Cloud-based software delivered via subscription (e.g., Microsoft 365, Slack). Eliminates need for on-premise installation.

SDLC Stages & Outputs

Order: Investigation: Feasibility study. Analysis: System requirements. Design: Technical specs (e.g., databases). Implementation: Functional system. Maintenance: Updates/optimizations.

Types of Feasibility

Economic: Cost vs. benefit (ROI). Technical: Can it be built with current resources? Operational: Will users adopt it? Legal/Regulatory: Compliance risks (e.g., GDPR).

Scope Creep

Uncontrolled expansion of a project's goals (e.g., adding features mid-development), often leading to delays/budget overruns.

Continuous Application Development

Frequent, iterative updates to software (e.g., monthly feature releases) via DevOps practices.

Agile Development

Iterative, collaborative approach emphasizing flexibility, customer feedback, and incremental deliverables (e.g., Scrum sprints).

Robotic Process Automation (RPA)

Software bots automating repetitive, rule-based tasks (e.g., invoice processing).

Accomplishes

Reduces human error and operational costs.

Low-Code/No-Code Applications

Platforms (e.g., Power Apps) enabling non-developers to build apps via drag-and-drop tools. Enables: Faster prototyping and democratizes app development.

Four IT Acquisition Strategies

Questions: Code Writing: Build custom, use low-code, or buy off-the-shelf? Payment Model: Purchase, lease, or SaaS? Location: On-premise, cloud, or hybrid? Origin: In-house, outsourced, or open-source?

Startup vs. Large Firm

Startup: Prefer SaaS/low-code for affordability. Corporation: May invest in custom solutions for scalability.

Turing Test

A test (by Alan Turing) where an AI's ability to mimic human responses is evaluated. If a human can't distinguish AI from a person, it 'passes.'

Machine Learning (ML) Benefits

Exceeds Humans In: Pattern recognition (e.g., fraud detection). Processing large datasets (e.g., genomics). Predictive analytics (e.g., demand forecasting).

Weak AI

Current AI (e.g., chatbots, Siri) is task-specific and lacks general intelligence ('strong' AI).

Deepfakes

AI-generated synthetic media (e.g., fake videos) using deep learning to impersonate real people.

Expert Systems

AI that emulates human expertise in a niche domain (e.g., IBM Watson for medical diagnosis).

Digital Twin

A virtual replica of a physical system (e.g., a factory) used for simulation and analysis.

GPS in Shipping Logistics

Tracks containers in real-time to optimize routes, reduce fuel costs, and prevent delays.

Recurrent Neural Network (RNN)

An AI model for sequential data (e.g., time series, speech) where outputs feed back into inputs.

Natural Language Processing (NLP) Example

Chatbots (e.g., Bank of America's Erica) use NLP to understand and respond to customer queries.

Intelligent Agent

Autonomous software that performs tasks (e.g., Amazon's recommendation algorithms).

ChatGPT Creator

OpenAI (launched in November 2022).

Chatbots in Finance

Automate customer service (e.g., balance inquiries, fraud alerts).

Predicting Customer Churn

Machine learning models analyze behavior (e.g., usage patterns) to flag at-risk customers.

Computer Vision

AI that interprets visual data (e.g., facial recognition, self-driving cars).

Deep Learning Characteristics

Uses neural networks with multiple layers. Requires massive datasets. Excels in unstructured data (e.g., images, text).

AI Pros & Cons

User Advantages: Personalization, 24/7 support. User Disadvantages: Privacy risks, bias in outputs. Business Advantages: Cost savings, efficiency. Business Disadvantages: High upfront costs, ethical dilemmas.