Trajectory
The step-by-step path an AI agent takes to complete a task, including states and actions.
Imagine an AI agent trying to achieve a goal, like finding a specific piece of information or playing a game. The trajectory is the complete sequence of steps it takes from the very beginning to the end. This includes every decision it makes and every action it performs along the way.
For example, if an AI agent is tasked with booking a flight, its trajectory would be the series of searches it performs, the flights it considers, the seats it picks, and the final confirmation. Understanding the trajectory helps developers see how the AI arrived at its solution and where it might have gone wrong.
An AI agent's trajectory is the sequence of states and actions it executes while pursuing a specific objective. This path represents the agent's behavior and decision-making process over time. Analyzing trajectories is crucial for understanding agent performance, debugging errors, and developing methods for agents to learn from past experiences, such as through reinforcement learning or imitation learning.
Recent advancements focus on enabling AI agents to learn from their past actions and outcomes. This involves analyzing the trajectories of successful and unsuccessful attempts to improve future performance. Tools like Google Gemini's ReasoningBank highlight the growing importance of understanding and learning from these step-by-step processes.
- ReasoningBank: Enabling agents to learn from experienceGoogle · Google Research Blog ·
- Google Gemini
- Google Gemini's ReasoningBank aims to improve AI agents by allowing them to learn from their past trajectories, both successes and failures.
- Path
- A path typically refers to a sequence of locations or nodes, whereas a tra
- Sequence
- A sequence is a general ordering of items, while a trajectory specific
- History
- History is a broad term for past events; a trajectory is a structured r
The term 'trajectory' in the context of AI and control systems originates from mathematics and physics, referring to the path of a moving object. In AI, it was adopted to describe the sequence of states and actions taken by an agent. Source ↗
Ask an AI assistant like ChatGPT or Google Gemini to write a short story about a robot exploring a new planet. Then, ask it to describe the robot's 'trajectory' or step-by-step journey, including its observations and actions. This will give you a sense of the sequence of events the AI generated.
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- PromptRole-play a student struggling with math for AI tutor simulationSimulate a student's math struggles and uncertainty for AI tutor evaluation. · 5 min
- TemplateReusable workflow templates for healthcare teamsTeams can create and share templates for common workflows like documentation or prior authorizations. · 15 min
- TemplateTemplate for building an AI tool tailored to your workA structured approach to defining and building a custom AI tool for specific business needs. · 15 min
- TemplateEvals CSV prompt list for agent skillsA CSV file template for listing prompts used in evaluating agent skills. · 20 min
- HeuristicDouble coding agent task completion with docs linksProviding coding agents with a documentation link significantly increases their task completion rate. · 2 min
Questions people ask
- What is an AI agent's trajectory?
- It's the complete sequence of steps, decisions, and actions an AI agent takes from start to finish while trying to achieve a goal.
- Why are trajectories important?
- They help developers understand how an AI works, debug problems, and train AI to perform better by learning from past experiences.
First seen on Subvolts Apr 21, 2026. Updated Sep 15, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.