The Agentic Service Designer
How AI Agent Teams, Service Blueprints-as-Code, and autonomous tooling are transforming service design from insight to operations
How AI Agent Teams, Service Blueprints-as-Code, and autonomous tooling are transforming service design from insight to operations
Read online at the book website · Download the PDF · Download the EPUB
Part I grounds the operating system. Parts II–IV walk the discipline end to end: research ops, mental models, five-lane blueprinting, the sixth agentic lane, orchestration, ideation, prototyping, recovery, quality governance. Part V is new in this edition: four dedicated platform chapters on the ASD MCP server (29 tools), the plugin host packs and Claude Code native plugin, BUZZ multi-agent engagements with Canvas indexing, and the DeepSeek Harness Studio workbench, each with real screenshots, exact commands and honest evidence boundaries. Part VI closes with SOPs-as-code, frontline enablement, Good Services audits, a complex fictional mastercase run across every surface, and the future of the practice.
BUZZ channel runs, Canvas round trips, DSH Studio sessions and CLI gate reviews: screenshots from bounded runs, captioned with what they prove and what they do not.
MCP JSON-RPC walkthroughs, host configurations for Claude Code / OpenCode / Codex, the ten-target pack matrix, persona-pack identity audits.
Meridian Bay Ferries: one fictional project across CLI, plugin, MCP, BUZZ and Studio: evidence chain, a Critic veto with remediation, human phase gates throughout.
TISDD, Kumar, Stickdorn, Downe, SDL and the Service Profit Chain, cited through 74 methods encoded as SOPs and served back out through the MCP server.
From sticky notes to living, autonomous service systems
One canonical practice, twelve compiled surfaces
Why agent teams fit service-dominant practice
Handoffs are artifacts, never chat
Two hundred interviews, one agent team
Quant and qual under one evidence roof
Closing the gap between belief and behaviour
Actors, exchanges and hidden dependencies
Journey maps that update themselves
The operational X-ray of a service
Where agents act, and how we gate them
One service, many channels, zero drift
Diverge wide, converge ruthlessly
Test the service before it exists
Designing for change and for failure
From projects to permanent capability
Twenty-nine tools over stdio
Install the whole practice into your coding agent
Ten personas, exact mentions, one Canvas
ASD inside a browser studio
Methods become executable, versioned, reviewed
The internal service that makes external ones possible
Lou Downe's bar, enforced by machine checks
One fictional project across every surface
What compounds when methods run
Complete reference of 74 methods and the artifact vocabulary
All 29 sd_* tools, arguments, configs and protocol facts
The knowledge corpus behind the methods
Ten role agents, 74 method skills, vendored MCP.
Stdio JSON-RPC; six verified bundles.