The Four Claude Entry Points for the CCAO-F Exam

Four surfaces behind one interface

Most people meet Claude as a chat box and never look further, and for a quick question that is exactly right. But the Claude Certified Associate - Foundations (CCAO-F) exam treats the choice of entry point as the first decision that shapes an entire session. Claude appears in one interface, yet that interface offers four distinct working surfaces, and each one is built for a different task pattern. Choosing the wrong surface does not usually produce a bad answer; it produces a good answer with far more effort than the task needed, session after session.

The four are Chat, Projects, Artifacts, and Research Mode. They differ in one thing above all: how much of the context you build today carries forward to tomorrow, and how much setup they ask for in return. Recognising them as four purpose-built tools, rather than one box with a few buttons, is the remember-level skill this knowledge point tests.

Chat and Projects: the two conversation surfaces

Chat is the default. It is a single conversation with no deliberate retained setup: you open it, type, and get an answer. A Chat is still saved to your history and can be reopened later, and Memory plus past-chat search can carry a few key facts into new sessions, so it is not strictly amnesiac. What Chat lacks is a Project's deliberate persistence -- standing instructions and a curated knowledge base that prime every new conversation automatically. That makes Chat ideal for one-off questions, quick tasks, and exploratory work where there is no plan to reuse a whole configured context.

Projects are the opposite by design. A Project is a persistent workspace that holds standing instructions, an uploaded knowledge base, and its own conversation history, so that every conversation you start inside it begins already primed with that context. Where Chat forgets, a Project remembers. The trade is setup cost: a Project takes a few minutes to build, which only pays off when the same background recurs across sessions. The detail of how that bundle is structured is covered in project anatomy, and the test for when the setup is worth it in the project worth-building test.

Artifacts and Research: output and investigation surfaces

Artifacts is not another conversation surface -- it is a rendered output pane. When the thing you are producing is itself the deliverable that someone will open and read as a standalone document, an Artifact renders it cleanly for review and sharing rather than leaving it buried in the chat thread. A common trap is to assume Artifacts is only for code; in practice it suits any reviewable deliverable, including a report or a structured document.

Research Mode drives multi-step, multi-source investigation. It is the surface for questions that cannot be trusted until Claude has gathered and synthesised across many sources. It is distinct from a quick current-information lookup, which web search inside Chat already handles. Knowing where that line sits is the subject of Research vs web search vs Artifacts.

What the CCAO-F exam trips candidates on

The exam tests two misreadings here. The first is treating Claude as a single undifferentiated chat box, so a candidate reaches for plain Chat regardless of the task pattern, missing that a recurring workstream belongs in a Project and a deep investigation belongs in Research Mode. The credited reading matches the surface to the task pattern before any prompt is written.

The second is assuming Artifacts is only for code. A report, a policy summary, or any structured document that a recipient will open and review is exactly what an Artifact is for. The word to anchor on is deliverable: if the output is the thing being handed over, it belongs in an Artifact regardless of whether it contains a line of code.

Common misreadings to avoid

How this shows up on the exam

Questions on this knowledge point describe a task and ask which entry point fits, or list the surfaces and ask you to match each to its purpose. The reliable approach is to read the task pattern first: is it one-off, recurring, a deliverable, or an investigation? That pattern names the surface.

This is the foundation for the rest of the domain. It leads directly into project anatomy, which unpacks what a Project actually holds, and pairs with the project worth-building test for deciding when the setup earns its keep. Getting the entry point right is what makes every later decision -- capability features, model tier, context management -- land on solid ground.