Apply & Reference

Enterprise AI Dictionary

Concise definitions with links into deeper Learn material.

Agent

A system that observes, reasons, and acts toward a goal under a runtime.

Coming Soon

Agent Harness

Execution runtime that governs context, tools, state, and authorization around an LLM.

Open →

Agent Runtime

The deterministic layer that owns sequencing, permissions, and recovery.

Coming Soon

Agent Loop

Observe → reason → act → evaluate until a stop condition.

Coming Soon

Agentic Workflow

A workflow that uses an LLM for some decisions, not unbounded autonomy.

Coming Soon

Context

The assembled workspace a model sees for a given turn.

Coming Soon

Context Engineering

Assembling, bounding, and governing what the model is allowed to see.

Open →

Memory

Persisted information retrieved later — not the current prompt.

Coming Soon

State

Durable workflow or session facts required to resume correctly.

Coming Soon

Tool

An external capability invoked with structured arguments.

Coming Soon

Capability

A permissioned, discoverable action a runtime may grant.

Coming Soon

RAG

Retrieval-augmented generation: fetch evidence, then generate.

Coming Soon

Reranker

A second-stage ranker that reorders retrieved candidates.

Coming Soon

Grounding

Tying generation to retrieved or deterministic evidence.

Coming Soon

MCP

A protocol for exposing tools and context to model clients.

Coming Soon

Eval

A measured check of quality, safety, or task success.

Coming Soon

Model Gateway

Shared entry point for model providers, routing, and policy.

Coming Soon

AI Gateway

Enterprise control plane in front of models, tools, and spend.

Coming Soon

Continuous Authorization

Re-evaluate access at the execution boundary, not once per login.

Coming Soon

Workload Identity

Identity for services and runtimes, independent of a human session.

Coming Soon

Delegation

Acting on behalf of another principal with a constrained grant.

Coming Soon

Trust Boundary

A line where identity, data, or control changes hands.

Coming Soon

Durable Execution

Work that survives crashes via checkpoints and replay.

Coming Soon

Human-in-the-Loop

A required human decision inside an otherwise automated path.

Coming Soon

Control Plane

Policy, identity, routing, and configuration — not the model loop.

Coming Soon

Execution Plane

Where tools actually run and side effects occur.

Coming Soon

Outcome Coordination

Commitments, verification, and settlement instead of fragile tool choreography.

Open →

Semantic Search

Retrieval by meaning using embeddings rather than exact terms.

Coming Soon

Vector Search

Nearest-neighbour lookup over embedding space.

Open →

Model Routing

Sending a request to the right model for cost, quality, or policy.

Coming Soon