What is AI?
AI is software that can process information, recognize patterns, make decisions, and generate useful output. Example: a model can summarize a failed Jenkins log and suggest the likely root cause.
Core concepts
Understand the real building blocks of AI and agents.
These are the terms you need to understand clearly before building anything serious. They are not just buzzwords. Each one matters in how AI systems behave and how you design reliable agents.
Learning note: This content is intended for learning and guidance only. If a concept or example does not fit your environment or goals, ignore it and keep the parts that help you.
What is a prompt?
A prompt is the instruction you give the model. It tells the AI the task, context, constraints, and how you want the answer structured.
What is a workflow?
A workflow is the sequence of steps an AI follows: read input → gather context → analyze → decide → return result. Example: inspect Terraform → find risky configuration → explain impact → suggest fix.
What is an LLM?
An LLM is a large language model. It predicts the next text based on learned patterns and can generate explanations, summaries, code, and structured output.
What is RAG?
RAG means Retrieval-Augmented Generation. It combines search and retrieval with AI generation so the model uses your docs, runbooks, or architecture files as context.
What is an agent?
An agent is an AI system with a goal. It can use tools, follow a workflow, and iterate until it solves the task or reaches a stop condition.
What is a skill?
A skill is a reusable instruction set for one job, such as AWS cost optimization, Kubernetes troubleshooting, or Jenkins failure review. It sets boundaries and expected output.
What is MCP?
MCP stands for Model Context Protocol. It is a standard interface for connecting an AI system to tools, files, APIs, and local services in a structured way.
What is a multi-agent workflow?
This is when one agent does one job and then calls another agent for the next stage. Example: a Research Agent gathers sources, a Summarizer Agent writes a brief, and a Validator Agent checks for mistakes.
In plain English
Beginner examples
A prompt is the instruction you give the AI. Example: “Review this Terraform file and list 3 possible security issues.”
A workflow is the step sequence. Example: read file → detect issue → explain risk → suggest fix.
An LLM is the model that generates human-like answers from the prompt and context.
An agent is a system that does a real job using a prompt, workflow, tools, and memory.
A skill is a reusable job-specific instruction set. Example: Kubernetes troubleshooting skill.
MCP is a way for the agent to connect to tools and external systems in a structured way.