Index

AI terms, capabilities, and workflow language.

Every term we use, defined plainly. Substrate, runtime, evals, governance, retrieval, workflow patterns, and closed-loop operations — for non-experts, start with Enterprise AI, RAG, MCP, and Evals.

New to AI? Start with these five terms:AI Agent Tokens RAG Hallucination MCP

B

Bad Thread Clusters Batch Inference Batch Pipeline Modernization Beyond model bake-offs Beyond prompt and program optimization

D

Dashboard as a living system Dashboard composition through language Data Catalog Data Quality Data Quality Gates Data substrate Detecting incidents in human↔agent conversations Document Extraction Drift Alerts Dynamic dashboard generation Dynamic runtime rendering

E

Embeddings Emerging trends Enterprise AI Escalation Policy Eval Case Capture Eval Dashboards Evals Evaluation Evaluation Harness Event Stream Ingestion Execution flow — ten stages External Memory

F

Fine-Tuning Future possibilities Future vision

G

Gate vs experiment Generated Code Tests Generative UI Governance Governed Datasets Guardrails Guardrails at every layer

H

Hallucination How extraction is run How it actually works How it actually works — the mentor–apprentice loop How it differs from a knowledge graph or lineage graph How it runs on AWS How optimization works How they are built How we cluster How we enforce it Human Approval Human In The Loop Human Review Queue Hybrid Retrieval

I

Inference Ingestion Contracts Intelligent caching systems Intent and entities — what the request is, and what it operates on Intent Routing Issue To Pr

K

Keep humans where they matter Knowledge Base Knowledge bases plugged in Knowledge Graph Knowledge Updates

L

Lakehouse Lakehouse Architecture LangChain LangGraph Latency Budgets Launch pattern Lineage Mapping LiteLLM LLM As Judge LLM Cost LLM-as-judge calibration LLM-Ready Knowledge Base llms.txt

M

MCP MCP Tool Registry Mechanics Memory architectures Model Fallback Strategy Model Gateway Model Routing Model routing per turn Model Routing Policy Multi Agent System Multi-client rendering

N

O

Observability Open questions we are studying Operational realities Operations Outcome extraction: did the user actually solve it?Owner Detection

P

Permissions in retrieval Persistent analytical memory Persistent conversational memory pgvector Prior art and adjacent work Private Inference Product Signal Production Agent Interfaces Promotion Gates Prompt and Model Diffs Prompt Caching Prompt Injection Prompt Model Diffs Prompt Policy Prompt-driven editing Pydantic Pydantic AI

Q

R

RAG Reflection and validation Regression Datasets Replaying the thread around the moment Reranking Reranking Policy Research notes Resolution Qa Responses API Retrieval is a system, not a database call Retrieval Readiness Retrieval-augmented generation Risk Review Root Cause Analysis Routing axes Runtime parameters

V

Vector Search Vercel AI SDK Voice Agents