Self-Evolving AI Agent Glossary
Plain-English definitions of AI agent terminology. Every term explains not just what it means, but how it differs from what people confuse it with — plus how self-evolving agents replace rented SaaS.
Core Concepts (14)
Agentic AI
AI that operates autonomously over extended tasks, making decisions without step-by-step human direction.
Read definition →AI Agent
An autonomous AI system that takes actions on your behalf to accomplish a goal.
Read definition →AI Agent Memory
The mechanisms by which an AI agent retains and recalls information across sessions and tasks.
Read definition →AI Copilot
An AI assistant embedded in a product that helps users complete tasks within that product's context.
Read definition →AI Guardrails
Constraints and safety mechanisms that limit what an AI agent can do to prevent unintended actions.
Read definition →AI Hallucination
When an AI model generates factually incorrect information with apparent confidence.
Read definition →AI Orchestration
The coordination and management of multiple AI agents or LLM calls within a system.
Read definition →AI Workflow
A predetermined sequence of steps with AI embedded at specific points — as opposed to an autonomous agent.
Read definition →Autonomous Agent
An AI agent that operates without human intervention to complete multi-step tasks.
Read definition →Generative AI
AI systems that generate new content — text, images, code, audio — rather than only classifying or analyzing input.
Read definition →Human-in-the-Loop (HITL)
An AI system design pattern where humans review or approve key agent decisions before they're executed.
Read definition →LLM (Large Language Model)
A deep learning model trained on large text corpora that can generate, understand, and reason with language.
Read definition →Multi-Agent System
An architecture where multiple specialized AI agents collaborate to complete complex tasks.
Read definition →Self-Evolving
An AI system that improves over time based on user feedback, patterns, and performance.
Read definition →Technical (22)
Agentic RAG
A RAG system where an AI agent controls the retrieval strategy adaptively, rather than following a fixed pipeline.
Read definition →Chain-of-Thought (CoT)
A prompting technique that asks an AI model to reason step-by-step before producing a final answer.
Read definition →Composio
An integration platform that provides 900+ OAuth-managed tool connections for AI agents.
Read definition →Context Window
The maximum amount of text an LLM can process in a single interaction.
Read definition →Embeddings
Numerical vector representations of text that encode semantic meaning.
Read definition →Few-Shot Prompting
Including a small number of input-output examples in a prompt to guide an AI model's behavior.
Read definition →Fine-Tuning
Additional training of a pre-trained LLM on domain-specific data to specialize its behavior.
Read definition →Function Calling
A model capability that allows LLMs to request calls to predefined functions with structured arguments.
Read definition →Inference
The process of running a trained AI model to generate output from a given input.
Read definition →LangChain
An open-source framework for building LLM-powered applications and AI agents in Python and JavaScript.
Read definition →MCP (Model Context Protocol)
Anthropic's open standard for connecting AI models to external tools and data sources.
Read definition →MCP Server
A process that exposes tools, resources, and prompts to MCP-compatible AI clients.
Read definition →OAuth
An open authorization standard that lets apps access user data from other services without sharing passwords.
Read definition →Prompt Engineering
The practice of designing inputs to AI models to reliably produce desired outputs.
Read definition →RAG (Retrieval-Augmented Generation)
A technique that grounds AI responses in retrieved documents rather than relying solely on training data.
Read definition →Semantic Search
Search that finds results by meaning rather than exact keyword matching.
Read definition →System Prompt
Instructions provided to an LLM before the conversation begins, setting its role, behavior, and constraints.
Read definition →Tool Call
A specific invocation of an external tool by an AI agent during task execution.
Read definition →Tool Use
An AI model's ability to call external functions, APIs, or services during reasoning.
Read definition →Vector Database
A database optimized for storing and searching high-dimensional vector embeddings.
Read definition →Webhook
A mechanism for one system to notify another system in real time when an event occurs.
Read definition →Zero-Shot Prompting
Asking an AI model to complete a task with no examples of the desired output.
Read definition →Business & Use Cases (5)
AI SDR
An AI agent that handles outbound sales development tasks — research, outreach, and follow-up — autonomously.
Read definition →AIOps
The use of AI and machine learning to automate and enhance IT operations, monitoring, and incident management.
Read definition →Knowledge Base (AI)
A collection of documents indexed for retrieval by an AI agent to answer questions accurately.
Read definition →Rented SaaS
Software delivered as a service through a recurring subscription, where you never own the software.
Read definition →Software Ownership
You own the code, data, and future of the software that runs your business.
Read definition →Ready to build self-evolving AI agents that own your software?
Shogo lets you deploy self-evolving AI agents that learn from your feedback — and gradually replace the rented SaaS your team depends on today.
Build Your First Agent Free →