# Red Finch Advisory > Plain definitions of the AI terms enterprise teams use, grouped by topic, with other names, words to avoid and files for AI tools. ## AI Lexicon (version 1.0, 2026-10-21) - [AI Lexicon as Markdown](https://redfinchadvisory.com/lexicon.md) - [AI Lexicon as JSON](https://redfinchadvisory.com/lexicon.json) ### What is AI - [AI](https://redfinchadvisory.com/lexicon/ai.md) - [Foundation model](https://redfinchadvisory.com/lexicon/foundation-model.md) - [Frontier model](https://redfinchadvisory.com/lexicon/frontier-model.md) - [Generative AI](https://redfinchadvisory.com/lexicon/generative-ai.md) - [Jagged frontier](https://redfinchadvisory.com/lexicon/jagged-frontier.md) - [LLM](https://redfinchadvisory.com/lexicon/llm.md) - [Machine learning](https://redfinchadvisory.com/lexicon/machine-learning.md) - [Model](https://redfinchadvisory.com/lexicon/model.md) - [Multimodal](https://redfinchadvisory.com/lexicon/multimodal.md) - [Open source AI](https://redfinchadvisory.com/lexicon/open-source-ai.md) - [Open weights](https://redfinchadvisory.com/lexicon/open-weights.md) - [Reasoning model](https://redfinchadvisory.com/lexicon/reasoning-model.md) - [Small language model](https://redfinchadvisory.com/lexicon/small-language-model.md) ### How models work - [Batch inference](https://redfinchadvisory.com/lexicon/batch-inference.md) - [Context window](https://redfinchadvisory.com/lexicon/context-window.md) - [Deterministic](https://redfinchadvisory.com/lexicon/deterministic.md) - [Embedding](https://redfinchadvisory.com/lexicon/embedding.md) - [Embedding model](https://redfinchadvisory.com/lexicon/embedding-model.md) - [Fine-tuning](https://redfinchadvisory.com/lexicon/fine-tuning.md) - [Inference](https://redfinchadvisory.com/lexicon/inference.md) - [Knowledge cutoff](https://redfinchadvisory.com/lexicon/knowledge-cutoff.md) - [Parameters](https://redfinchadvisory.com/lexicon/parameters.md) - [Post-training](https://redfinchadvisory.com/lexicon/post-training.md) - [Pretraining](https://redfinchadvisory.com/lexicon/pretraining.md) - [Probabilistic](https://redfinchadvisory.com/lexicon/probabilistic.md) - [Reasoning effort](https://redfinchadvisory.com/lexicon/reasoning-effort.md) - [Token](https://redfinchadvisory.com/lexicon/token.md) - [Training](https://redfinchadvisory.com/lexicon/training.md) ### Ways of working - [Agentic](https://redfinchadvisory.com/lexicon/agentic.md) - [Agentic coding](https://redfinchadvisory.com/lexicon/agentic-coding.md) - [Agentic workflow](https://redfinchadvisory.com/lexicon/agentic-workflow.md) - [AI-assisted](https://redfinchadvisory.com/lexicon/ai-assisted.md) - [AI-assisted coding](https://redfinchadvisory.com/lexicon/ai-assisted-coding.md) - [AI-assisted workflow](https://redfinchadvisory.com/lexicon/ai-assisted-workflow.md) - [AI-generated](https://redfinchadvisory.com/lexicon/ai-generated.md) - [Assistant](https://redfinchadvisory.com/lexicon/assistant.md) - [Custom assistant](https://redfinchadvisory.com/lexicon/custom-assistant.md) - [Software factory](https://redfinchadvisory.com/lexicon/software-factory.md) - [Vibe coding](https://redfinchadvisory.com/lexicon/vibe-coding.md) - [Workflow automation](https://redfinchadvisory.com/lexicon/workflow-automation.md) ### Agents - [Agent](https://redfinchadvisory.com/lexicon/agent.md) - [Background agent](https://redfinchadvisory.com/lexicon/background-agent.md) - [Coding agent](https://redfinchadvisory.com/lexicon/coding-agent.md) - [Enterprise agent](https://redfinchadvisory.com/lexicon/enterprise-agent.md) - [Internal agent](https://redfinchadvisory.com/lexicon/internal-agent.md) - [Multi-agent system](https://redfinchadvisory.com/lexicon/multi-agent-system.md) - [Orchestration](https://redfinchadvisory.com/lexicon/orchestration.md) - [Personal agent](https://redfinchadvisory.com/lexicon/personal-agent.md) - [Router agent](https://redfinchadvisory.com/lexicon/router-agent.md) - [Subagent](https://redfinchadvisory.com/lexicon/subagent.md) ### Autonomy and identity - [Acts as itself](https://redfinchadvisory.com/lexicon/acts-as-itself.md) - [Acts as you](https://redfinchadvisory.com/lexicon/acts-as-you.md) - [Acts for you](https://redfinchadvisory.com/lexicon/acts-for-you.md) - [Advisor](https://redfinchadvisory.com/lexicon/advisor.md) - [Agent identity](https://redfinchadvisory.com/lexicon/agent-identity.md) - [Autonomous](https://redfinchadvisory.com/lexicon/autonomous.md) - [Human oversight](https://redfinchadvisory.com/lexicon/human-oversight.md) - [Human-in-the-loop](https://redfinchadvisory.com/lexicon/human-in-the-loop.md) - [Human-on-the-loop](https://redfinchadvisory.com/lexicon/human-on-the-loop.md) - [Human-out-of-the-loop](https://redfinchadvisory.com/lexicon/human-out-of-the-loop.md) - [Kill switch](https://redfinchadvisory.com/lexicon/kill-switch.md) - [Least privilege for AI](https://redfinchadvisory.com/lexicon/least-privilege-for-ai.md) - [Levels of autonomy](https://redfinchadvisory.com/lexicon/levels-of-autonomy.md) - [Task runner](https://redfinchadvisory.com/lexicon/task-runner.md) ### Instructions - [Context](https://redfinchadvisory.com/lexicon/context.md) - [Context engineering](https://redfinchadvisory.com/lexicon/context-engineering.md) - [Few-shot prompting](https://redfinchadvisory.com/lexicon/few-shot-prompting.md) - [Memory](https://redfinchadvisory.com/lexicon/memory.md) - [Prompt](https://redfinchadvisory.com/lexicon/prompt.md) - [Prompt chaining](https://redfinchadvisory.com/lexicon/prompt-chaining.md) - [Prompt engineering](https://redfinchadvisory.com/lexicon/prompt-engineering.md) - [Prompt template](https://redfinchadvisory.com/lexicon/prompt-template.md) - [Role prompting](https://redfinchadvisory.com/lexicon/role-prompting.md) - [Session](https://redfinchadvisory.com/lexicon/session.md) ### Connections and systems - [A2A](https://redfinchadvisory.com/lexicon/a2a.md) - [Agent gateway](https://redfinchadvisory.com/lexicon/agent-gateway.md) - [Agent harness](https://redfinchadvisory.com/lexicon/agent-harness.md) - [Agent loop](https://redfinchadvisory.com/lexicon/agent-loop.md) - [Agent skill](https://redfinchadvisory.com/lexicon/agent-skill.md) - [AI operations](https://redfinchadvisory.com/lexicon/ai-operations.md) - [Client](https://redfinchadvisory.com/lexicon/client.md) - [Code execution](https://redfinchadvisory.com/lexicon/code-execution.md) - [Computer use](https://redfinchadvisory.com/lexicon/computer-use.md) - [Connector](https://redfinchadvisory.com/lexicon/connector.md) - [Grounding](https://redfinchadvisory.com/lexicon/grounding.md) - [Interface](https://redfinchadvisory.com/lexicon/interface.md) - [Knowledge base](https://redfinchadvisory.com/lexicon/knowledge-base.md) - [LLM gateway](https://redfinchadvisory.com/lexicon/llm-gateway.md) - [Loop engineering](https://redfinchadvisory.com/lexicon/loop-engineering.md) - [MCP server](https://redfinchadvisory.com/lexicon/mcp-server.md) - [Model Context Protocol](https://redfinchadvisory.com/lexicon/model-context-protocol.md) - [Model routing](https://redfinchadvisory.com/lexicon/model-routing.md) - [Observability](https://redfinchadvisory.com/lexicon/observability.md) - [RAG](https://redfinchadvisory.com/lexicon/rag.md) - [Sandbox](https://redfinchadvisory.com/lexicon/sandbox.md) - [Stack](https://redfinchadvisory.com/lexicon/stack.md) - [System prompt](https://redfinchadvisory.com/lexicon/system-prompt.md) - [Tool](https://redfinchadvisory.com/lexicon/tool.md) - [Vector database](https://redfinchadvisory.com/lexicon/vector-database.md) ### Evaluation - [A/B test](https://redfinchadvisory.com/lexicon/a-b-test.md) - [Benchmark](https://redfinchadvisory.com/lexicon/benchmark.md) - [Eval](https://redfinchadvisory.com/lexicon/eval.md) - [Human-as-judge](https://redfinchadvisory.com/lexicon/human-as-judge.md) - [LLM-as-judge](https://redfinchadvisory.com/lexicon/llm-as-judge.md) - [Production monitoring](https://redfinchadvisory.com/lexicon/production-monitoring.md) - [Red teaming](https://redfinchadvisory.com/lexicon/red-teaming.md) - [Regression eval](https://redfinchadvisory.com/lexicon/regression-eval.md) - [Side-by-side eval](https://redfinchadvisory.com/lexicon/side-by-side-eval.md) - [Verified test set](https://redfinchadvisory.com/lexicon/verified-test-set.md) ### Risks - [AI safety](https://redfinchadvisory.com/lexicon/ai-safety.md) - [AI security](https://redfinchadvisory.com/lexicon/ai-security.md) - [AI slop](https://redfinchadvisory.com/lexicon/ai-slop.md) - [Alignment](https://redfinchadvisory.com/lexicon/alignment.md) - [Bias](https://redfinchadvisory.com/lexicon/bias.md) - [Data exfiltration](https://redfinchadvisory.com/lexicon/data-exfiltration.md) - [Deepfake](https://redfinchadvisory.com/lexicon/deepfake.md) - [Deterministic guardrail](https://redfinchadvisory.com/lexicon/deterministic-guardrail.md) - [Guardrail](https://redfinchadvisory.com/lexicon/guardrail.md) - [Hallucination](https://redfinchadvisory.com/lexicon/hallucination.md) - [Indirect prompt injection](https://redfinchadvisory.com/lexicon/indirect-prompt-injection.md) - [Input guardrail](https://redfinchadvisory.com/lexicon/input-guardrail.md) - [Jailbreak](https://redfinchadvisory.com/lexicon/jailbreak.md) - [Output guardrail](https://redfinchadvisory.com/lexicon/output-guardrail.md) - [PII redaction](https://redfinchadvisory.com/lexicon/pii-redaction.md) - [Probabilistic guardrail](https://redfinchadvisory.com/lexicon/probabilistic-guardrail.md) - [Prompt injection](https://redfinchadvisory.com/lexicon/prompt-injection.md) - [Reward hacking](https://redfinchadvisory.com/lexicon/reward-hacking.md) - [Sycophancy](https://redfinchadvisory.com/lexicon/sycophancy.md) ### Oversight - [Agent sprawl](https://redfinchadvisory.com/lexicon/agent-sprawl.md) - [AI ethics](https://redfinchadvisory.com/lexicon/ai-ethics.md) - [AI governance](https://redfinchadvisory.com/lexicon/ai-governance.md) - [AI governance committee](https://redfinchadvisory.com/lexicon/ai-governance-committee.md) - [AI inventory](https://redfinchadvisory.com/lexicon/ai-inventory.md) - [Approved tools](https://redfinchadvisory.com/lexicon/approved-tools.md) - [Deployer](https://redfinchadvisory.com/lexicon/deployer.md) - [Explainability](https://redfinchadvisory.com/lexicon/explainability.md) - [Fine-tuning on our data](https://redfinchadvisory.com/lexicon/fine-tuning-on-our-data.md) - [High-risk AI](https://redfinchadvisory.com/lexicon/high-risk-ai.md) - [How AI uses our data](https://redfinchadvisory.com/lexicon/how-ai-uses-our-data.md) - [In-context use](https://redfinchadvisory.com/lexicon/in-context-use.md) - [ISO/IEC 42001](https://redfinchadvisory.com/lexicon/iso-iec-42001.md) - [Model risk management](https://redfinchadvisory.com/lexicon/model-risk-management.md) - [Provider](https://redfinchadvisory.com/lexicon/provider.md) - [Responsible AI](https://redfinchadvisory.com/lexicon/responsible-ai.md) - [Retrieval from our data](https://redfinchadvisory.com/lexicon/retrieval-from-our-data.md) - [Shadow AI](https://redfinchadvisory.com/lexicon/shadow-ai.md) - [Vendor training](https://redfinchadvisory.com/lexicon/vendor-training.md) ### People and roles - [Agent manager](https://redfinchadvisory.com/lexicon/agent-manager.md) - [AI brain fry](https://redfinchadvisory.com/lexicon/ai-brain-fry.md) - [AI center of excellence](https://redfinchadvisory.com/lexicon/ai-center-of-excellence.md) - [AI champion](https://redfinchadvisory.com/lexicon/ai-champion.md) - [AI literacy](https://redfinchadvisory.com/lexicon/ai-literacy.md) - [AI product manager](https://redfinchadvisory.com/lexicon/ai-product-manager.md) - [AI-pilled](https://redfinchadvisory.com/lexicon/ai-pilled.md) - [Chief AI Officer](https://redfinchadvisory.com/lexicon/chief-ai-officer.md) - [Citizen developer](https://redfinchadvisory.com/lexicon/citizen-developer.md) - [Cognitive debt](https://redfinchadvisory.com/lexicon/cognitive-debt.md) - [Entry-level gap](https://redfinchadvisory.com/lexicon/entry-level-gap.md) - [Evals engineer](https://redfinchadvisory.com/lexicon/evals-engineer.md) - [External FDE](https://redfinchadvisory.com/lexicon/external-fde.md) - [Forward deployed engineer](https://redfinchadvisory.com/lexicon/forward-deployed-engineer.md) - [Full-stack builder](https://redfinchadvisory.com/lexicon/full-stack-builder.md) - [Internal FDE](https://redfinchadvisory.com/lexicon/internal-fde.md) - [Meat proxy](https://redfinchadvisory.com/lexicon/meat-proxy.md) - [Product engineer](https://redfinchadvisory.com/lexicon/product-engineer.md) - [The great flattening](https://redfinchadvisory.com/lexicon/the-great-flattening.md) - [Tokenmaxxing](https://redfinchadvisory.com/lexicon/tokenmaxxing.md) - [Use-case owner](https://redfinchadvisory.com/lexicon/use-case-owner.md) ### Value case - [AI washing](https://redfinchadvisory.com/lexicon/ai-washing.md) - [Cost per task](https://redfinchadvisory.com/lexicon/cost-per-task.md) - [Cost per token](https://redfinchadvisory.com/lexicon/cost-per-token.md) - [GEO (generative engine optimization)](https://redfinchadvisory.com/lexicon/geo.md) - [Moat](https://redfinchadvisory.com/lexicon/moat.md) - [Model-agnostic](https://redfinchadvisory.com/lexicon/model-agnostic.md) - [MVP](https://redfinchadvisory.com/lexicon/mvp.md) - [Outcome-based pricing](https://redfinchadvisory.com/lexicon/outcome-based-pricing.md) - [Per-seat pricing](https://redfinchadvisory.com/lexicon/per-seat-pricing.md) - [Pilot](https://redfinchadvisory.com/lexicon/pilot.md) - [Production](https://redfinchadvisory.com/lexicon/production.md) - [Proof of concept (POC)](https://redfinchadvisory.com/lexicon/proof-of-concept.md) - [Sovereign AI](https://redfinchadvisory.com/lexicon/sovereign-ai.md) - [Usage-based pricing](https://redfinchadvisory.com/lexicon/usage-based-pricing.md) - [Wrapper](https://redfinchadvisory.com/lexicon/wrapper.md)