Procedural knowledge
Knowledge of how to perform a task, often expressed as a sequence of steps or rules. It's about doing, not just knowing facts.
Procedural knowledge is knowing how to do something. Think of it like following a recipe to bake a cake or giving directions to a friend. It's a set of instructions or steps that lead to a specific outcome.
For example, when an AI agent is given the procedural knowledge to "write an email," it knows the steps involved: open a new email, address it, write a subject line, compose the body, and send. This is different from knowing facts, like "Paris is the capital of France."
Procedural knowledge refers to the "how-to" aspect of intelligence, distinct from declarative knowledge (facts and figures). In AI, it's often encoded as algorithms, state machines, or learned policies that dictate a sequence of actions to achieve a goal. This can be explicitly programmed or implicitly learned through reinforcement learning or imitation learning.
Representations can range from simple if-then rules to complex neural network policies. Tradeoffs involve the flexibility of learned procedures versus the reliability and interpretability of explicitly defined ones. For agents, procedural knowledge is crucial for executing tasks in dynamic environments.
Recent advancements in AI agents, like those from Anthropic, highlight the need for AI to perform tasks. Procedural knowledge is essential for these agents to understand and execute sequences of actions required to interact with the real world or complex digital systems.
- Equipping agents for the real world with Agent SkillsAnthropic · Anthropic Engineering ·
- Anthropic Claude
- Anthropic's Agent Skills framework likely relies on agents possessing procedural knowledge to perform tasks like file management or executing specific workflows.
- Declarative knowledge
- This is knowledge of facts and information, like knowing
- Machine learning models
- While models can learn procedural knowledge, the term i
- Algorithms
- Algorithms are a common way to represent procedural knowledge, but p
The concept of procedural knowledge originates from cognitive psychology, distinguishing it from declarative knowledge. In AI, it's fundamental to the development of intelligent agents and expert systems.
Ask an AI assistant like ChatGPT or Google Gemini to "explain how to tie a shoelace" or "give me step-by-step instructions to make a peanut butter sandwich." Observe how it breaks down the task into a sequence of actions.
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's the difference between procedural and declarative knowledge?
- Declarative knowledge is knowing 'what' (facts, concepts), while procedural knowledge is knowing 'how' (skills, tasks, steps).
- How do AI agents use procedural knowledge?
- AI agents use procedural knowledge to perform tasks by following a sequence of steps, enabling them to interact with environments and achieve goals.
First seen on Subvolts Apr 15, 2026. Updated Sep 14, 2026. Written by an AI model from the items above and reviewed against them weekly; report an error.