# AshnaAI AI glossary

> Short definitions of AI, agents, and automation terms used across AshnaAI.

Hub: https://www.ashna.ai/ai-glossary
Last updated: 2026-08-14

## Terms

- [Agentic AI](https://www.ashna.ai/ai-glossary/agentic-ai): Autonomous AI systems that plan, reason, and act across tools with minimal human input.
- [AI Agents](https://www.ashna.ai/ai-glossary/ai-agents): Software entities that break goals into tasks and execute them using models, data, and APIs.
- [Artificial Intelligence (AI)](https://www.ashna.ai/ai-glossary/artificial-intelligence): Computer systems designed to perform tasks that typically require human intelligence.
- [Augmentation](https://www.ashna.ai/ai-glossary/augmentation): Enriching models with external context, tools, or data for more accurate outputs.
- [Benchmark](https://www.ashna.ai/ai-glossary/benchmark): Standardized tests used to compare model performance on tasks like reasoning or coding.
- [Chain of Thought (CoT)](https://www.ashna.ai/ai-glossary/chain-of-thought): A prompting technique where models reason step-by-step before giving a final answer.
- [Chunking](https://www.ashna.ai/ai-glossary/chunking): Splitting large documents into smaller pieces for efficient retrieval and RAG.
- [Confidence Score](https://www.ashna.ai/ai-glossary/confidence-score): A signal of how certain a model is about a prediction or classification.
- [Context Window](https://www.ashna.ai/ai-glossary/context-window): The maximum amount of text a model can process in a single request.
- [Conversational AI](https://www.ashna.ai/ai-glossary/conversational-ai): Systems that interact through natural language across chat, voice, and messaging.
- [Dense Retrieval](https://www.ashna.ai/ai-glossary/dense-retrieval): Semantic search using embeddings to find meaningfully similar content.
- [Embedding](https://www.ashna.ai/ai-glossary/embedding): Numerical vector representations that capture semantic meaning of text or data.
- [Fine-Tuning](https://www.ashna.ai/ai-glossary/fine-tuning): Training a base model further on domain-specific data for specialized behavior.
- [Foundation Model](https://www.ashna.ai/ai-glossary/foundation-model): Large general-purpose models adaptable to many downstream tasks.
- [Generative AI](https://www.ashna.ai/ai-glossary/generative-ai): AI that creates new text, images, code, or audio from learned patterns.
- [GPT (Generative Pre-trained Transformer)](https://www.ashna.ai/ai-glossary/gpt-generative-pre-trained-transformer): A family of transformer-based language models used for text generation and agents.
- [Grounding](https://www.ashna.ai/ai-glossary/grounding): Anchoring model outputs to trusted sources to reduce hallucinations.
- [Guardrails](https://www.ashna.ai/ai-glossary/guardrails): Policies and controls that keep AI outputs safe, compliant, and on-brand.
- [Hallucination](https://www.ashna.ai/ai-glossary/hallucination): When a model produces confident but incorrect or unsupported information.
- [Human in the Loop](https://www.ashna.ai/ai-glossary/human-in-the-loop): Keeping people involved in approvals, corrections, or escalations.
- [Hybrid Search](https://www.ashna.ai/ai-glossary/hybrid-search): Combining keyword and semantic search for better retrieval quality.
- [In-Context Learning](https://www.ashna.ai/ai-glossary/in-context-learning): Teaching models new tasks through examples in the prompt without retraining.
- [Large Language Model (LLM)](https://www.ashna.ai/ai-glossary/large-language-model): Neural networks trained on vast text corpora to understand and generate language.
- [Multi-Agent Orchestration](https://www.ashna.ai/ai-glossary/multi-agent-orchestration): Coordinating multiple specialized agents to complete complex workflows.
- [No-Code AI](https://www.ashna.ai/ai-glossary/no-code-ai): Building AI applications without traditional programming skills.
- [Prompt Engineering](https://www.ashna.ai/ai-glossary/prompt-engineering): Designing instructions and examples that steer model behavior reliably.
- [Retrieval-Augmented Generation (RAG)](https://www.ashna.ai/ai-glossary/retrieval-augmented-generation): Fetching relevant documents before generating an answer for better accuracy.
- [Token](https://www.ashna.ai/ai-glossary/token): The basic units models use to process text; affects cost and context limits.
- [Vector Database](https://www.ashna.ai/ai-glossary/vector-database): Storage optimized for similarity search over high-dimensional embeddings.
- [Workflow Automation](https://www.ashna.ai/ai-glossary/workflow-automation): Using AI to trigger and complete multi-step business processes.