TRANSLATION KEY // YOUR WORLD ⇄ COPILOT STUDIO // AB-620
Decryption complete /// vocabulary aligned

The Rosetta Stone

Every Copilot Studio noun, mapped to the concept you already carry from building agents on opencode and wiring MCP by hand. Read left→right to translate; watch the false friends where the mapping lies.

Your world / general agent-dev Copilot Studio term

Core agent

the agent / assistantAgent

Copilot Studio's agent bundles model + instructions + knowledge + tools + topics + triggers into one object. (Older docs say "copilot" or "bot" — same thing.)

system promptInstructions

Plain-text box on the agent's Overview page. Type / to reference a tool, topic, variable, or Power Fx expression inline — that's how you wire behavior to components.

choosing the modelModel choice / Foundry catalogue

You can now pick the underlying model (including Claude Sonnet 4.5 / Opus 4.5), or pull a model from the Foundry model catalogue for custom prompts.

the router / plannerOrchestration

The big one. Two modes — generative vs classic. See the keystone at the bottom; don't skim it.

Tools & integration

tools / function callingTools

Umbrella term — broader than function-calling. In Copilot Studio "tools" covers MCP tools, connectors, REST API tools, agent flows, and prompts. Anything the agent can invoke.

MCP server connectionMCP tool

Same protocol you run today. Add it as a tool and point it at the server. Basically free points for you.

an API call / HTTP requestCustom connector · Send HTTP request · REST API tool

Three wrappers for the same idea. Custom connector = reusable, auth-managed definition. Send HTTP request = raw one-off. REST API tool = expose an endpoint to the agent.

prebuilt SDK integrationsConnectors

Power Platform connectors (1000+ prebuilt to SaaS apps) and Copilot / Graph connectors (pull org content in as knowledge).

agent controls a browser/UIComputer use

Agent takes actions in a UI on the user's behalf. Know when it's the right tool vs an API call — that's the exam angle.

Knowledge & retrieval

RAG / retrievalKnowledge sources + generative answers node

Attach docs/sites as knowledge; the generative answers node answers from them instead of the base model. It's RAG with a GUI.

vector store / indexAzure AI Search

The grown-up grounding backend for unstructured docs. Often surfaced via Foundry.

querying a structured datasetFabric data agent

The structured-data counterpart to doc-RAG. Docs → Azure AI Search; analytical/semantic-model data → Fabric data agent; single live record → a connector.

Logic, state & data

scripted/guardrailed dialogTopic (nodes on a canvas)

A deterministic conversation path built from nodes. This is the un-LLM part — hardcoded flow for things that must be predictable. Optional trigger phrases fire it in classic mode.

a background job / workflow scriptAgent flow

Power Automate under the hood. Structured automation behind the chat — supports approvals, input/output params, error handling.

a human approval gateHuman-in-the-loop

An approval step inside an agent flow. Person confirms before the flow proceeds.

inline expressions / glue logicPower Fx

Excel-formula-flavored, not a general language. Low ceiling — read it, write simple ones, don't expect Python.

variables / context / stateVariables (global vs topic scope)

Scope is tested: global = whole agent; topic = local to one conversation path.

your databaseDataverse

Microsoft's managed data backend (tables/rows). Where agent + solution data lives.

rich/structured message UIAdaptive cards

JSON-defined UI cards for responses. Learn the JSON shape.

Multi-agent

sub-agents / in-system delegationConnected / child agents

Delegation within Copilot Studio. Your parent agent calls a specialized child.

cross-vendor agent interopA2A protocol

Hand-off across platforms/vendors. The connected-vs-A2A line (same platform vs cross-platform) is a recurring exam question.

calling an externally-hosted agentIntegrate a Foundry / existing agent

Bring a Foundry-hosted or already-built agent into your solution as a collaborator.

Deploy & operate

deploy targetsChannels

Teams, a website, M365 Copilot, custom — anything on Azure Bot Service.

env separation (dev/prod)Environments + Solutions

A solution is the deployable package (managed vs unmanaged); environments isolate dev/test/prod.

CI/CD promotionPower Platform Pipelines

Promote a solution dev→test→prod without manual rebuild.

env vars / secrets / configEnvironment variables

So endpoints/keys aren't hardcoded across environments.

observability / tracingApplication Insights

Telemetry for the deployed agent. Familiar concept, Azure branding.

evals / test suiteTest set + agent evaluation

Build cases, pick an eval method, review results. Domain 3 — the sleeper.

auth (OAuth, tokens)Connections + identity strategy

Connectors carry their own connection/auth. "Identity strategy" = the planning-level version (who authenticates, end-user vs agent).

safety layer / content policyResponsible AI + DLP policies

Guardrails/moderation/groundedness + Data Loss Prevention for governance.

False friends — where the mapping lies

"Tools" is broader than function-calling

In your head a tool ≈ a function the model can call. In Copilot Studio tools also include connectors, agent flows, and even prompts. When a question says "add a tool," it may mean any of those — read the scenario.

A "topic" is NOT a tool

Topics are scripted dialog paths, not callable functions. Closest analog is a hardcoded, deterministic conversation branch. If something must be predictable/compliant, it's a topic; if it's open-ended reasoning, that's the model + generative answers.

"Orchestration" has a specific dual meaning

It does not mean your orchestration-framework sense. It's a binary setting on the agent: generative vs classic. Answering an "orchestration" question with framework intuition will burn you.

Power Fx is not a programming language

It looks like code but it's Excel-formula-flavored with a low ceiling. Don't reach for loops/classes — the exam expects you to read and write simple expressions, nothing more.

Agent flow ≠ topic

A topic is conversation (what the agent says). An agent flow is background automation (what the agent does — Power Automate). Questions test whether you reach for the right one.

Keystone concept — learn this cold

Generative vs Classic orchestration

This is the one concept with no clean 1:1 in your world, and it shows up everywhere on the exam.

Generative orchestration

The agent decides at runtime which combination of topics, tools, and knowledge to use for a given request — driven by the names/descriptions you write for each. Flexible, LLM-planned.

Classic orchestration

The user's message is matched to a topic's trigger phrases via NLU, and that topic's scripted path runs. Deterministic, author-controlled.

Mental model: generative = you describe capabilities and trust the model to compose them; classic = you wire explicit intent→flow routes. Modern agents lean generative; classic is the fallback for tight, predictable control. Know which one a scenario is describing and you've answered half its follow-ups.