The Shape of This Track
A free lesson from Swipe AI · AI Agents & Automation, level 1 · updated October 2026
By the Swipe team, founded by Colin · how we make lessons
Here's the full roadmap. Level 2 names the actual repeating cycle every agent runs on: perceive, decide, act. Level 3 goes deeper into tools specifically — not just that they exist, but how an agent decides which one to reach for and when.
Levels 4 and 5 cover what makes multi-step behavior actually work: planning a goal out into steps ahead of time, and memory — what an agent keeps track of within a single run versus across separate runs.
L2-L3: The loop -> perceive/decide/act, choosing and using tools L4-L5: What makes it work -> planning multi-step tasks, memory L6-L7: Scaling up -> multiple agents working together, how much autonomy to grant L8-L9: Staying safe -> failure modes, human-in-the-loop design L10: Capstone -> tied together, bridges to AI Ethics & Safety
The second half scales up and turns practical: Level 6 covers multiple agents working together, Level 7 covers how much autonomy to actually grant an agent, and Levels 8-9 cover what happens when agents go wrong and how to design safe human oversight — which connects directly to the accountability questions this app's AI Ethics & Safety track covers. Level 10 ties the whole arc together.
What does Level 2 of this track establish?
Levels 8 and 9 cover failure modes and ____-in-the-loop design for safe oversight.
This track's later levels on safety connect directly to the accountability themes covered in AI Ethics & Safety.
More Than a Chatbot
10 questions · pass with 60% — just like in the app.
What fundamentally distinguishes an AI agent from a plain single-response chat interaction?
An AI agent requires a fundamentally different, non-language-model technology than plain chat interactions.
An agent can keep going in a ____ until its actual goal is reached, instead of stopping after one output.
How does this track relate to Building With AI APIs' function-calling content?
This track assumes no prior knowledge of function calling or tool use whatsoever.
Building With AI APIs taught what a single tool call looks like; this track teaches how to string many together with real ____ in between.
What does Level 2 of this track establish?
This track's safety-related levels connect directly to the accountability themes covered in AI Ethics & Safety.
Levels 8 and 9 cover failure modes and ____-in-the-loop design.
What is the overall arc of this track's ten levels?
Free ai tools
Keep your streak going in Swipe AI
Swipe through every level of AI Agents & Automation on your iPhone — with XP, stars, streaks, a personal advisor and the full track beyond these free levels.
Get the Swipe apps →