# Glossary

Key terms used throughout the playbook, in alphabetical order.

| Term | Definition |
|------|-----------|
| Agent | An AI system given a specific role, scope, and set of instructions to perform a defined task. In this playbook, "agent" can mean a simple role-prompted AI chat or a more sophisticated autonomous AI system. |
| Archetype | A learning style profile that describes how you naturally approach new skills. The four archetypes in this playbook ([Tinkerer](/archetypes/tinkerer/), [Planner](/archetypes/planner/), [Strategist](/archetypes/strategist/), [Social Learner](/archetypes/social-learner/)) are derived from AI Skills Quiz results. |
| Cross-Domain Reframing | The skill of borrowing AI techniques, frameworks, or prompting patterns from one field and adapting them to work in another. One of the five [pillars of AI fluency](/pillars/cross-domain-reframing/). |
| Ethical Prompting | The practice of using AI responsibly: verifying outputs, checking for errors, considering audience impact, and maintaining appropriate human judgment. One of the five [pillars of AI fluency](/pillars/ethical-prompting/). |
| Fluency | In this playbook, fluency means practical proficiency with AI — the ability to use AI tools effectively, critically, and ethically across different contexts. Not just knowing what AI can do, but being skilled at making it work well. |
| Handoff | The point where output from one AI session or agent becomes input for the next. Managing handoffs — what information to pass, what to leave out, and how to structure the transfer — is a core agent collaboration skill. |
| Insight Synthesis | The skill of extracting structured, actionable meaning from AI-generated content — distinguishing signal from noise, triangulating across multiple outputs, and forming your own judgment. One of the five [pillars of AI fluency](/pillars/insight-synthesis/). |
| Multi-Agent System | A setup where multiple AI agents (each with different roles and contexts) work on parts of the same problem, with outputs coordinated by a human or automated orchestrator. |
| Pathway | A guided route through the playbook's exercises, curated for a specific quiz result profile. Pathways recommend which exercises to do, in what order, based on your strengths and gaps. |
| Pillar | One of five core skill areas that make up AI fluency: [Insight Synthesis](/pillars/insight-synthesis/), [Workflow Automation](/pillars/workflow-automation/), [Cross-Domain Reframing](/pillars/cross-domain-reframing/), [Agent Collaboration](/pillars/agent-collaboration/), and [Ethical Prompting & Judgment](/pillars/ethical-prompting/). |
| Prompt Chain | A sequence of AI prompts where each step's output feeds into the next step's input. The simplest form of an AI workflow. |
| Quality Gate | A checkpoint in an AI workflow where output is evaluated before proceeding to the next step. Prevents errors from propagating through a pipeline. |
| Tokenization | The process by which AI models break text into smaller chunks called *tokens* before processing. Tokens can be whole words, parts of words, or punctuation. For example, "tokenization" might become "token" + "ization." Token counts determine input/output limits in AI tools. See [How AI Actually Works](/concepts/how-ai-actually-works/) for more context. |

| Workflow Automation | The skill of turning repeatable tasks into systematic AI processes — from reusable prompt templates to multi-step pipelines with quality gates. One of the five [pillars of AI fluency](/pillars/workflow-automation/). |