aug 30, 2025
👋 i am walid boulanouar, co‑founder of ay automate. i write this newsletter to break down complex concepts in ai and automation in a clear, actionable way. my mission is to help you understand how ai agents can transform operations, save time, and drive growth.
this newsletter is reader‑supported. to receive new posts and join our community of forward‑thinking businesses, consider becoming a subscriber.
the topic that has been dominating headlines in the ai agent world for the past months context engineering. in this article i outline practical notes from building agentic systems over the past few years. we will look at all types of context that agentic systems rely on and the challenges that come with managing it.
for engineers who have been building ai agents, context engineering is a new name but not a new practice it is how we guided agents to perform the work in the first place. the “old school” prompt engineering is a subset of context engineering.
prompt engineering a subset of context engineering
a simplified agentic system topology
in simple terms, it is a topology of llm calls connected via different patterns. each output of an llm node influences the downstream system.
the quality of an agent is only as good as the context you pass into prompts at each step. bigger context windows are not a silver bullet. pushing too much data creates issues like:
context poisoning a hallucination makes it into context
context distraction context overwhelms training
context confusion superfluous context influences the response
context clash parts of the context disagree
context engineering aims to provide the minimal, focused context at each step so the agent can do its job.
below are common context types and challenges.
different types of context
