# The Blueprint > The Blueprint is a field guide for business leaders building organisations where people and AI agents share the work: org charts of roles, role charters, plain-English playbooks, commitment ledgers, sponsors and Agentic Resources, across 20 chapters grounded in real cases and current research. A Field Guide to Building Organisations of People and Agents - [The Blueprint](https://garethsprice.github.io/agentic-org-blueprint-book/index.md): Overview of The Blueprint, a field guide for business leaders building organisations where people and AI agents share the work, covering org charts of roles, role charters, plain-English playbooks, commitment ledgers, sponsors and Agentic Resources across 20 chapters. ## Part I: The Shift - [Introduction: The Day the Bots Multiplied](https://garethsprice.github.io/agentic-org-blueprint-book/introduction.md): Introduction to The Blueprint: why multiplying AI agents is an org design problem, the author's route from a founder who couldn't delegate to a CTO delegating to agents, and a map of the book. - [Chapter 1: You Already Know How to Do This](https://garethsprice.github.io/agentic-org-blueprint-book/1-you-already-know-how.md): Why executives adopt AI agents fastest, how they delegate the wrong layer (the how, or other people's work), the Shape, Steer and Do model of work, and the structural fixes that keep agents inside the org chart. - [Chapter 2: Systems Run the Business](https://garethsprice.github.io/agentic-org-blueprint-book/2-systems-run-the-business.md): Gerber's and Deming's systems thinking rebuilt for AI agents: most failures belong to the system, model dependency as the new key-person risk, the agent that quietly does six jobs, and the cost of being the human glue. - [Chapter 3: Draw the Finished Org Chart First](https://garethsprice.github.io/agentic-org-blueprint-book/3-draw-the-finished-org-chart.md): How to design an organisation of people and agents: draw the finished org chart of roles before filling it, write the Blueprint well enough to stand the organisation up again tomorrow, and size the chart for the company you are becoming. ## Part II: The Destination - [Chapter 4: Know What It's For](https://garethsprice.github.io/agentic-org-blueprint-book/4-know-what-its-for.md): Chapter 4 of The Blueprint: defining the Aim, the Destination (including the mix of people and agents), a small paired Scorecard, and a purpose Story that agents can act on. ## Part III: The Org Chart - [Chapter 5: Hire the Role, Not the Model](https://garethsprice.github.io/agentic-org-blueprint-book/5-hire-the-role-not-the-model.md): Chapter 5 of The Blueprint: a role is not an agent. One occupant can hold several chartered roles; the number of agents is a separate staffing decision; apply the role coherence test. - [Chapter 6: One Boss, Few Reports](https://garethsprice.github.io/agentic-org-blueprint-book/6-one-boss-few-reports.md): Chapter 6 of The Blueprint: hierarchy versus swarm for AI agents. Coordinators contain errors, one role writes, each role has one boss, spans stay small, and peer teams work only when tests do the managing. - [Chapter 7: Every Agent Needs a Sponsor](https://garethsprice.github.io/agentic-org-blueprint-book/7-every-agent-needs-a-sponsor.md): Why every AI agent needs a named human sponsor accountable for its purpose and lifecycle, how to assign sponsors by function, and why agents need their own identity and signature. - [Chapter 8: What and Why, Never How](https://garethsprice.github.io/agentic-org-blueprint-book/8-what-and-why-never-how.md): How to brief AI agents with what and why rather than how, why over-specified instructions hold capable models back, and how a verification bar replaces step-by-step procedure. ## Part IV: Writing It Down - [Chapter 9: One Page Per Job](https://garethsprice.github.io/agentic-org-blueprint-book/9-one-page-per-job.md): How to write a one-page Role Charter for a human or AI agent role: outcome line, owned and served work, quantity, quality and behaviour standards, and acceptance agreed before work starts. - [Chapter 10: Bet on Plain English](https://garethsprice.github.io/agentic-org-blueprint-book/10-bet-on-plain-english.md): Chapter 10 argues that plain-English playbooks beat visual workflow builders for directing agents, drawing on VHS, PHP, the Bitter Lesson and CorralData's own adoption record. - [Chapter 11: Lock In Only What Can't Go Wrong](https://garethsprice.github.io/agentic-org-blueprint-book/11-lock-in-only-what-cant-go-wrong.md): Chapter 11 sets out three tiers for agent instructions (English playbooks, scripts for facts and arithmetic, hard rules only at irreversible points) and how to review lock-ins as models improve. - [Chapter 12: Everything They Need on Day One](https://garethsprice.github.io/agentic-org-blueprint-book/12-everything-they-need-on-day-one.md): Chapter 12 defines the Role Pack, the lean package of charter, playbooks, reasons, boundaries and sources of truth that any occupant needs, and how to stop it drifting. ## Part V: Running the Machine - [Chapter 13: Track What's Owed, Not What's Done](https://garethsprice.github.io/agentic-org-blueprint-book/13-track-whats-owed.md): How to run a Commitment Ledger for people and agents: record what each role owes and to whom, close items only on evidence, flag slips early, separate scheduled from event-driven work, and give every escalation an owner. - [Chapter 14: Humans at the Irreversible Points](https://garethsprice.github.io/agentic-org-blueprint-book/14-humans-at-the-irreversible-points.md): Why most human approval of AI agents is theatre (people accept 93% of prompts), and how to place people only at irreversible actions, demand evidence agents cannot write, and promote agents through draft, ask and act tiers on evidence. - [Chapter 15: Work in Public](https://garethsprice.github.io/agentic-org-blueprint-book/15-work-in-public.md): How to stop work going missing at handoffs between people and AI agents: requests in public channels not DMs, least-privilege access requests, bots that can be reached, handoffs that link to state, and working in public as the way agents learn. ## Part VI: Getting Better - [Chapter 16: The Model Isn't the Problem](https://garethsprice.github.io/agentic-org-blueprint-book/16-the-model-isnt-the-problem.md): How to turn every recurring AI agent failure into a system fix: Deming's view of the system, the MAST failure taxonomy, the Friction Review, and evals built from real failures. - [Chapter 17: Agentic Resources](https://garethsprice.github.io/agentic-org-blueprint-book/17-agentic-resources.md): Agentic Resources, a term coined by Gareth Price for HR for agents: the org-chart function that hires, onboards, reviews, promotes, reconciles, re-staffs and retires AI agents. ## Part VII: The Human Side - [Chapter 18: When the Work Moves](https://garethsprice.github.io/agentic-org-blueprint-book/18-when-the-work-moves.md): A guide for leaders moving work from people to AI agents: make it an explicit decision, be open, fair and firm, let domain experts own the agents, and build new career paths. - [Chapter 19: From Doing to Delegating](https://garethsprice.github.io/agentic-org-blueprint-book/19-from-doing-to-delegating.md): The individual contributor's side of delegating to agents: precise but cautious adoption, personal agents, the renegotiated deal of outcomes in and verification out, and why reviewing other people's agent output burns people out. - [Chapter 20: Your Job Now](https://garethsprice.github.io/agentic-org-blueprint-book/20-your-job-now.md): The closing chapter of The Blueprint: the leader's job in an organisation of people and agents is Shape work, consistency, holding standards, deciding at irreversible points, keeping the Blueprint current and acting while small teams have the advantage. ## Appendices - [Appendix A: The First Thirty Days](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-a-first-thirty-days.md): A thirty-day, week-by-week build order for leaders: Aim and Scorecard, an agent census, the finished org chart of roles, a first Role Charter and Playbook, a Commitment Ledger, a first Friction Review, earned autonomy and an Agentic Resources owner. - [Appendix B: Templates](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-b-templates.md): Original copy-paste templates for running an organisation of people and agents: Role Charter, tiered Playbook, Role Pack index, Commitment Ledger with status vocabulary, Friction Review, and Agentic Resources onboarding and offboarding checklist, each with a worked mini example. - [Appendix C: A Worked Example](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-c-worked-example.md): A fictional composite company's customer success function built with The Blueprint method: Aim and Scorecard, an org chart of roles, two Role Charters, a tiered Playbook, a Role Pack, Commitment Ledger entries, a Friction Review, and the sponsor and Agentic Resources setup. - [Appendix D: For Your Technical Team](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-d-for-your-technical-team.md): Technical appendix mapping the book's org-design vocabulary (roles, charters, Role Packs, three lock-in tiers, Commitment Ledger, sponsors, acceptance) to agent definitions, skills, hooks, deny rules, progress files, durable execution, Entra Agent ID, MCP 2026-07-28, A2A 1.0, evals and observability. - [Appendix E: Glossary](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-e-glossary.md): Glossary of The Blueprint's terms (Role Charter, Role Pack, Commitment Ledger, Friction Review, sponsor, Agentic Resources, three lock-in tiers) and common agent terms (agent, skill, MCP, A2A, eval, harness, hook, context window), each linked to its chapter. - [Appendix F: Recommended Reading](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-f-recommended-reading.md): Annotated reading list for The Blueprint: Gerber, Grove, Drucker, Deming, Gawande, Ohno, Team Topologies and Conway on organisation; Marquet, McChrystal and Wiseman on delegation; Sutton, Anthropic, Google Research, Cognition and protocol specifications on building with agents; and Applied Alchemy. - [Appendix G: Your AI Org Design Advisor](https://garethsprice.github.io/agentic-org-blueprint-book/appendix-g-ai-org-design-advisor.md): Copy-pasteable SKILL.md that turns Claude, ChatGPT or Grok into an AI org design advisor grounded in The Blueprint: interviews the leader, drafts an org chart of roles and Role Charters, finds duplicate bots and missing sponsors, and proposes Friction Reviews. ## Key Concepts - [What Is Agentic Resources?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-agentic-resources.md): Definition of Agentic Resources, a term coined by Gareth Price in The Blueprint for HR for AI agents: the org-chart function that owns the agent workforce from hiring to retirement, how it differs from sponsorship, how to start this week, and how it relates to earlier Agent Resources proposals. - [What Is an AI Agent Org Chart?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-an-ai-agent-org-chart.md): Definition of an AI agent org chart as used in Gareth Price's The Blueprint: a chart of roles, each with an outcome line and a human owner, sized for the company you are becoming, with people and agents as occupants, one boss per role and small spans. Includes an example and how to draw one this week. - [What Is a Role Charter?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-a-role-charter.md): Definition of a Role Charter, the one-page role specification in Gareth Price's The Blueprint for people and AI agents alike: outcome line, owns and serves, quantity, quality and behaviour standards, authority, acceptance agreed in advance, and how it differs from a job description or a system prompt. - [What Is an Agent Sponsor?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-an-agent-sponsor.md): Definition of an agent sponsor as used in Gareth Price's The Blueprint: the one named human accountable for an AI agent's purpose, access, performance and retirement, why function owners make the best sponsors, and how a sponsor differs from an agent's owner, administrator or manager. - [What Is a Commitment Ledger?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-a-commitment-ledger.md): Definition of a Commitment Ledger as used in Gareth Price's The Blueprint: a running record of what each role in a team of people and AI agents has promised, closed only on evidence, kept outside the worker, with slips raised early, and how it differs from a to-do list or a ticket tracker. - [What Is Shape, Steer, Do?](https://garethsprice.github.io/agentic-org-blueprint-book/what-is-shape-steer-do.md): Definition of Shape, Steer, Do, the model of work in Gareth Price's The Blueprint adapted from Michael Gerber's entrepreneur, manager and technician: Shape decides what the organisation becomes, Steer makes results predictable, Do is the work itself. Explains why agents pull leaders back into Do and how to delegate the right layer. - [By the Numbers](https://garethsprice.github.io/agentic-org-blueprint-book/by-the-numbers.md): By the Numbers, from The Blueprint by Gareth Price: field data from CorralData's team of people and agents (bot counts, shadow periods, failures) and the research figures the book cites, each linked to its chapter and source. ## Optional - [About the Author](https://garethsprice.github.io/agentic-org-blueprint-book/author.md): About Gareth Price -- CTO at CorralData, author of Applied Alchemy, former technologist at The New York Times and Codecademy, who runs a mixed team of people and AI agents. - [How This Book Was Built](https://garethsprice.github.io/agentic-org-blueprint-book/process.md): How The Blueprint's first draft was produced by a team of AI agents organised with the book's own method -- roles, briefs, sponsors and verification -- and what the human author still had to do. - [Acknowledgements](https://garethsprice.github.io/agentic-org-blueprint-book/acknowledgements.md): Acknowledgements for The Blueprint, including its debt to Michael E. Gerber's E-Myth books and the colleagues and agents whose work supplied its cases. - [What's New](https://garethsprice.github.io/agentic-org-blueprint-book/updates.md): Dated change log for The Blueprint, a field guide to organisations of people and AI agents, listing new pages, revised guidance and refreshed research.