Cognitive Orchestration
A twelve-week course for builders, creators, operators, educators, and teams who want to work with AI more deliberately, not just more frequently.
14
Weekly modules
6
Blueprint areas
AI
Work systems
Advanced · Flagship operating course
What you will be able to do more clearly
A complete professional operating model for AI-enabled knowledge work: problem anatomy, delegation, workflow, instruction, evaluation, memory, orchestration, bounded autonomy, and judgment.
- Level
- Advanced
- Lessons
- 21
- Format
- 14 weeks
- Access
- Facilitated cohort
Who it's for
Builders, creators, operators, founders, educators, researchers, and institutional teams
Syllabus
21 lessons, in order
A deliberate sequence — each lesson answers a real problem and sets up the next. Lesson 1 is free to preview; the rest unlock when you enroll.
- 11. Start Here: From Memory to AgentsStart here — the lay of the land before the machinery: how this course fits together, why a persistent agent amplifies whatever structure you give it, and where it all begins, with memory and the system prompt.Free preview20-30 min
- 22. Prompt EngineeringOnce a model is in front of you, the quality of what you get back depends heavily on how you ask. Prompt engineering is the practice of structuring requests — instructions, examples, and format — so the model reliably produces what you need.Free preview20-30 min
- 33. Retrieval-Augmented Generation (RAG)A model only knows what it was trained on — frozen at a cutoff date, with no access to your private documents. Retrieval-Augmented Generation, introduced by Lewis et al. (2020), fixes this by fetching relevant text at query time and feeding it into the prompt.Free preview20-30 min
- 44. Context EngineeringAs soon as systems run over many turns with tools and retrieved data, the prompt is no longer one message — it is a crowded, finite context window. Context engineering is the discipline of curating that window: choosing the right tokens to have present at each step, and managing everything else.Free preview20-30 min
- 55. Model Context Protocol (MCP)Retrieval and context get the right information into a model. Tools let a model do things — query a database, send an email, open a file. The problem was that every app wired up tools its own way. The Model Context Protocol (MCP), open-sourced by Anthropic in late 2024, standardises that wiring.Free preview20-30 min
- 66. Workflow Automation (n8n, Make, Zapier)Not every problem needs a free-roaming agent. Often you want repeatable, auditable steps — and that is the home of no-/low-code automation platforms. They are where most teams first put LLMs to productive work.Free preview20-30 min
- 77. Agentic AI & AI AgentsAt the top of the stack sits agentic AI: systems where the model does not just answer — it directs its own process, deciding which tools to call and in what order to reach a goal. This is the home ground of the runtimes this course is built on: Hermes Agent runs as a long-lived agent on your own server — persistent memory, autonomous skill creation, subagent delegation, and a Kanban board where named agents collaborate — while OpenClaw is a personal agent you message on WhatsApp, Telegram, or Slack that controls your computer and works on its own schedule, even overnight. Anthropic's distinction is the cleanest mental model for reasoning about both: a workflow orchestrates LLMs through predefined code paths; an agent lets the model dynamically control how it accomplishes the task. With a workflow, you own the plumbing; with an agent, the model owns it.Free preview20-30 min
- 88. First Principles and Problem AnatomyDiagnose the real structure of work — outcome, constraint, evidence, judgment — before deciding whether a persistent agent is warranted at all.Free preview45–60 min
- 9. The Mindshift and Security EnvironmentsShift from chat-tool habits to runtime thinking, and set the trust, credential, and execution boundaries before an agent touches your environment.45–60 min
- 10. Calibrated Delegation and Headless IntegrationsDecide what to keep, assist, or delegate, and wire agents into headless tools and channels without over-granting their reach.45–60 min
- 11. Procedural Memory and Interface DesignPackage reusable skills with the right instructions, scripts, and limits so the right capability appears at the right moment.45–60 min
- 12. Knowledge Infrastructure and Graph RetrievalDesign the memory stack — reference, procedural, and graph knowledge — and retrieve it well with RAG and graph methods.45–60 min
- 13. Evaluation, Verification, and EvidenceBuild evaluation and verification into agentic workflows so outputs are backed by evidence rather than assertion.45–60 min
- 14. Orchestration Patterns: Ephemeral and DurableChoose between workflows and agents, and between ephemeral and durable orchestration, for each kind of task.45–60 min
- 15. Bounded Autonomy and Least-Privilege ActionGrant the least authority that still gets the work done, with hard boundaries and human checkpoints around consequential action.45–60 min
- 16. Evolutionary Prompt OptimizationTreat prompts and instructions as artifacts you measure and improve systematically rather than tweak by feel.45–60 min
- 17. Trajectory Harvesting and Reinforcement LearningCapture agent trajectories and reflective traces to turn lived experience into durable improvement.45–60 min
- 18. External Environment ControlLet agents act on real environments — browser, shell, computer use — through tools and protocols, safely scoped.45–60 min
- 19. Human Judgment and Cognitive StewardshipKeep human judgment where consequence and accountability live, guarding against automation bias and de-skilling.45–60 min
- 20. Personal Operating Systems and Team AdoptionTurn an individual operating model into shared team practice without losing clarity, safety, or accountability.45–60 min
- 21. The Master Blueprint and CapstoneSynthesize the six layers into one operating blueprint and apply it end-to-end in a capstone.45–60 min
Outcomes
What should stay with you
Not short-term inspiration — a stronger way to interpret, reason, govern, anticipate, and act.
- Diagnose work structurally before delegating to AI.
- Assign tasks across people, models, tools, memory, and review loops.
- Evaluate outputs and build safeguards for AI-enabled workflows.
- Produce a Cognitive Orchestration Blueprint for a real workflow.
What's included
Built for individuals, cohorts, and institutions
Use the course as a guided reading experience, a facilitated cohort, an internal training program, or a partner academy module.
Included
Twelve lesson readings
A complete sequence from problem anatomy to final blueprint.
Included
Blueprint workbook
Templates for workflow, delegation, verification, memory, and governance design.
Included
Facilitator guide
Discussion prompts for teams and cohort delivery.
Included
Delivery sequence
Email and cohort rhythm for guided participation.
Pathways
Where this course leads next
Deepen the work through another program, apply it with your team, or move toward a fellowship or public contribution.
Recommended next step
Trust, Risk, and Governance
Apply orchestration thinking to systems that can act across tools, workflows, and environments.
Explore next step
Recommended next step
Institutional partnerships
Bring the course into a team, partner academy, or internal capability program.
Explore next step
Ready to start Cognitive Orchestration?
Read lesson 1 free, then register interest to join the next facilitated cohort and unlock all 21 lessons.