Glossary

Architecture

• Software architecture is the fundamental organization of a software system
• Blueprint that defines the system’s high-level structure
• Encompasses system’s components, the relationships among them, and the principles guiding its design and evolution
• Key characteristics i.e. the ones architecture is concerned about: scalability, maintainability, reliability, availability, security, performance, testability, extensibility

Design

• The process of conceptualizing and planning how a software system will work to satisfy its requirements
• Focuses on how the system’s components, modules, functions, algorithms, and data structures will be implemented
• Detailed planning stage before the coding begins

Architecture (A) vs Design (D)

• Analogy:
  • A: architectural plan for a city; deals with the what are the major parts and why are they structures this way
  • D: detailed blueprints for an individual building; deals with the how are we building the individual parts
• Focus:
  • A: high-level structure, components and their relationships; high-level design (HLD)
  • D: detailed implementation, internal workings of individual modules, components, classes, and algorithms; low-level design (LLD)
• Concerns:
  • A: non-functional requirements like scalability, availability, security, performance, maintainability
  • D: algorithms, data structures, component interactions, functions, and interfaces
• Patterns:
  • A: Architectural Patterns e.g. microservices, monolithic, event-driven, service-oriented
  • D: Design Patterns e.g. Factory, Singleton, Observer
• Change impact:
  • A: decisions are costly and difficult ti change once implemented as they form the system’s foundation
  • D: decisions are easier to change later as they are localized

Agentic AI

• Agentic AI is a concept.
• The key concept here is Agentic - from the concept of “agency”, which can be thought of as “autonomy”.
• Autonomy in decision making, planning, and taking action.
• Properties of Agentic AI systems:
  • Autonomy: operate independently
  • Goal oriented: work toward predefined goals
  • Perceive: gather data from multiple sources, such as sensors, databases, and digital interfaces
  • Reason: LLMs as reasoning engines
  • Act: task execution by means of tools integration
  • Learn: continuous improvement through feedback loop, or “data flywheel”, where data generated from its interactions is fed into the systems to enhance models

AI Agent

• AI Agent is an autonomous, goal-oriented system or application.
• It can be said to be an implementation of the Agentic AI concept.

Process Automation

• PA is the use of software and technology to automate repetitive, manual tasks and entire functions within a business process
• aims to reduce human intervention, increase efficiency, and minimize errors across formalized, rule-based workflows

Digital Process Automation

• Focuses on transforming and automating end-to-end business processes

Task Automation

• automates a single, simple, repetitive activity
• uses simple tools, scripts, or macros
• e.g. auto-replying to a specific email type

Robotic Process Automation

• focused on user interface
• automates tasks by mimicking human actions on a computer screen
• replicate human interactions for high-volume, repetitive work
• uses software robots (bots)
• e.g. a bot logging into a system and copying data from a spreadsheet