CoachDiscoverMapCompareSavedAI LearningAI WeeklyAbout
← All guides
Guide 01 · Foundations

What AI in SharePoint actually is

A worker agent that writes and runs code in a sandbox — not a chatbot that searches your files. Understand the model before you trust the math.

Walk away with
Worker agent ≠ RAG
Walk away with
Code runs in a sandbox
Walk away with
Tool calls drive the flow

AI in SharePoint is a worker agent

AI in SharePoint behaves like a co-worker who can read files, write code, and produce output — not like a chatbot that answers from memory. When you hand it a prompt, it discovers files, decides on a workflow, executes that workflow in steps, and writes the result back to the library as a real file.

The model architects the flow and writes the code. The code does the work. You get a file back, not a paragraph of guesses.

This is much closer to Claude Code or a co-work agent than it is to a chat experience. The agent starts a job knowing nothing about your data, then unpacks, discovers, writes code, and executes — all internally — to deliver the artifact you asked for.

People assume AI in SharePoint is "Copilot with extra steps." It is not. Copilot is a chat/RAG experience grounded in search; AI in SharePoint is a worker agent that runs code against your data.

Why this is not Copilot or RAG

Copilot is a Retrieval-Augmented Generation (RAG) system: it does a search, pulls some chunks of content, and asks the LLM to reason over what it found. That works for chat-style answers but breaks down on big numeric work — the LLM ends up estimating, sampling, or saying "based on what I saw."

RAG asks the LLM to be the calculator. AI in SharePoint asks the LLM to write the calculator, then runs it.

This split is why Copilot agents do not run inside SharePoint the same way. Copilot agents are RAG agents that customize search and grounding. AI in SharePoint is a different product class entirely — a worker agent that gets work done by writing and executing code, not by predicting answers.

"If I add enough Copilot agents I'll get the same result." You won't. RAG cannot do real math at scale — it will estimate and confabulate when the data is too large for the context window.

The sandbox is a trust boundary

Every line of code AI in SharePoint writes runs inside a secure sandbox. The sandbox is the trust boundary: data flows in, the agent writes code, the code runs against that data, and only the resulting artifact (an HTML file, a JSON blob, an image) flows out.

The sandbox can manipulate the data it is given and produce output. It cannot call external services or modify other content on your behalf.

This is why you cannot hand AI in SharePoint a script to run — that constraint is a security feature, not a limitation. The agent generates code dynamically inside the sandbox; users never inject code. Anthropic's Skills allow user-provided code; AI in SharePoint deliberately does not.

"No user code means no dynamic code." Wrong. The system still generates and runs code dynamically — it just generates it itself, inside the boundary, on data the user supplied.

Tool calls are the workflow

Under the hood, AI in SharePoint runs a chain of tool calls. The model picks the right tool, hands it parameters, and uses the result to decide what to call next. A typical job: find filesread filesrun code in sandboxgenerate HTMLcreate file.

Think Unix pipes. Each tool does one thing well. The model wires them together based on the prompt.

The "read files" step looks familiar if you know Unix — it is essentially cat, returning the content of files. The "run code" step is where pandas-style analytics happen. The "create file" step is what writes the final report back into the document library.

"The LLM is reading the files." It is not. The model sees metadata and structure; the actual bytes of the files stay in tool storage on the server and are only touched by code the agent wrote.
Flashcards — Foundations
01 · Foundations
AI in SharePoint
tap to reveal →
A worker agent that writes and runs code in a sandbox against SharePoint files. Different from Copilot, which is a chat/RAG experience grounded in search.
← tap to flip back
01 · Foundations
Worker Agent
tap to reveal →
An AI agent that delivers work artifacts — files, reports — by writing and executing code, rather than predicting answers from a trained model.
← tap to flip back
01 · Foundations
RAG
tap to reveal →
Retrieval-Augmented Generation. The LLM searches a corpus, retrieves chunks, and reasons over them. Powers Copilot; cannot do precise math at large scale.
← tap to flip back
01 · Foundations
Sandbox
tap to reveal →
A secure execution boundary where agent-generated code runs against data the user supplied. Cannot call external services or mutate other content.
← tap to flip back
01 · Foundations
Tool Call
tap to reveal →
A single step the agent invokes — find files, read files, run code, create file. The model chains tool calls to build a workflow.
← tap to flip back
Guide 01 · Foundations · AI in SharePoint All flashcards →