Clicked Gallery

What is context engineering?

Highlighted from a real engineering doc. Explained by Clicked.

Used in a sentence

Engineering Notes · AI Systems

Most of the reliability gains came from context engineering rather than from any change to the prompt.

The reader highlighted one word in the docs. Clicked explained the technical term “context engineering” in simple terms:

Explained in three depths

Same facts, different vibe — Slang mode 😎

The Clicked way

●○○

Overview

Context engineering is the work of deciding everything a model sees before it answers. The team building the system sets the strategy, and the code around the model applies it each time a question arrives. Four things fill that space: the standing instructions, facts fetched from your documents, the conversation so far, and the tools the model may use. Prompt engineering is how you word the question. Context engineering is what sits in front of the model while it reads that question.
●○○

Overview

Somebody has to decide what the model is looking at when your question lands, and that job is context engineering. Four slots get filled: the standing rules, whatever got pulled from the files, the chat so far, and the tools it may use. Fill those well and even a clumsy question gets a decent answer. Fill them badly and no amount of careful phrasing will save the reply. 😎

A quick take — often all you need.

●●○

Detail

Context engineering is the work of assembling what a model reads before it answers. Which sources get searched, how much of the earlier conversation survives, what gets summarised once it grows too long, which tools the model is even told exist: all of that is decided by somebody. Ask a support bot about your refund policy and it invents one, because nobody arranged for the policy to be there. You could paste the policy in yourself, though nobody does that for every customer, so the system has to assemble it instead. Why does this matter more than a sharper prompt? A prompt shapes one answer, while a system running twenty steps needs the right material at every one of them. The constraint is that the context window is finite, so more is not better and adding everything you own buries the part that mattered. When an AI feature works in testing and fails in production, this assembly is usually where it broke.
●●○

Detail

Context engineering is everything that gets loaded into a model's head before your question arrives. A support bot asked when an order shipped will invent a date with total conviction, having never been shown the shipping records. You can hand those over by hand every single time, or build something that does it for you, and building that is context engineering. Documents are only part of it. How much of the old chat survives, what gets trimmed when it runs long, which tools the thing is allowed to touch, all of it counts. Why not write a sharper instruction instead? Instructions steer one reply, and anything doing real work chews through twenty steps that each want their own fresh input. Piling on more does not help either, because space runs out and the useful page sinks under the other four hundred. 😎

Want more? One click digs deeper.

●●●

Analogy

A surgeon about to operate. How the request was worded matters very little at this point. What matters is whether the scans on the screen belong to this patient, whether the chart lists the right allergy, and whether the instruments laid out are the ones this operation needs. Hand a brilliant surgeon the wrong file and the operation goes badly, however good the surgeon is.
Deepest level reached.
●●●

Analogy

A washing machine engineer turning up at your door. They already know how washing machines work, that part is long done. Whether the visit goes anywhere depends on being sent with the manual for your model, the right part in the van, and a tool that fits your machine. Same engineer, wrong manual, and you get an immaculate repair to a machine that lives somewhere else. 😎
Deepest level reached.

Unfamiliar concept? A real-world example makes it click — fresh analogies on tap.

AI explanations may contain errors · Not professional advice

Formal definition — The same term, explained the usual way

Context engineering is the practice of curating the non-prompt material supplied to a language model at inference time, comprising system instructions, retrieved documents, conversation state and tool definitions. It is distinguished from prompt engineering, which concerns the formulation of the instruction itself, and it subsumes retrieval-augmented generation as one of its mechanisms. Its constraints are the finite context window and the degradation in accuracy that results from supplying irrelevant material.

Want Clicked to explain terms like “context engineering” directly in your browser — including on PDFs?

Add to Chrome — Free

50 free Explanations · No credit card required