The complete map

Full-stack agent-first design.

Agent-first design isn't one discipline — it's five layers moving together. Vision, ownership, processes, technology, and the AUX trust layer that decides whether any of it gets adopted.

This page is the index. It takes the five-layer model from the 5-Step Playbook and maps every relevant auxfirst resource — guides, essays, frameworks, and definitions — onto the layer it actually serves. Start where your gap is. The trust layer wraps all of them.

LAYER 01
Strategy & Design
What do we want to become?
LAYER 02
Ownership
Who is accountable?
LAYER 03
Processes
What is the work, really?
LAYER 04
Technology
What stack serves it?
LAYER 05
AUX · Trust
What earns adoption?
01
Strategy & Design · Vision

What do we want to become?

Before a dollar goes on tooling, leadership has to agree on what "agentic" means for you — where agents create leverage, where humans stay in the loop, and how far you'll let software act. This is the layer that sets the north-star and the trust posture everything else points at.

EngagementExecutive SeminarAlign founders and executives on what agentic actually means — and what to decide next. ManifestoThe AUX ManifestoFrom interface design to relationship design — the core argument of the agentic era. EssayWhat Is Agentic UX?The AUX Start Pack — manifesto, patterns, heuristics, trust architecture, evolution curve. PlaybookBecoming Agentic — 5 StepsThe transformation field guide this map is built on, turn-by-turn per layer. PracticeAI Product ExperienceTurning a SaaS from a static tool into an adaptive partner people trust. ReferenceKnowledge BaseThe full auxfirst dictionary — every AUX concept, pillar, and term, cross-linked. GlossaryAI-Native ProductsProducts designed for agents from the ground up — not retrofitted. GlossaryAgent-Led GrowthWhen the agent becomes the channel — a new motion for adoption.
02
Ownership · Accountability

Who is accountable — and who owns trust?

Agentic initiatives die in the gap between "everyone's excited" and "no one's accountable." An agent that acts inside your business is a product with a user base, and it needs real ownership — including a named owner for trust and adoption, not just one for whether it works.

FrameworkAgency Brains DesignDefining, naming, and governing the AI "brains" an organization runs. FrameworkBrains Design — Implementation AtlasThe turn-by-turn build sequence for standing up governed agent brains. PillarThe Agentic Family OfficeBounded agents for wealth operations — governance, entities, advisors, stewardship. GlossaryHuman-in-the-Loop SystemsWhere a human approves before the agent acts — the default for high stakes. GlossaryHuman-on-the-Loop SystemsSupervisory control — the agent acts, the human monitors and can override. GlossaryMulti-Agent Experience DesignWho acted, who delegated, who owns the outcome — clarity across many agents. GlossaryTrust in AI SystemsThe four-stage trust architecture — functional, contextual, judgment, advocacy.
03
Processes · The Work

What is the work, really?

The fastest way to fail is to point an agent at a process you've never actually mapped. This layer separates the documented workflow from the real one that runs on exceptions, then designs the delegation: what the agent owns, what stays human, where control changes hands.

PracticeAgent Process DesignTurning a workflow into a defensible agentic service — two lenses, ten stages. EngagementBlueprint SprintFrom process reality to a build-ready agent design — maps, blueprints, specs. PillarAgent-First Design — The GuideThe master discipline — process, patterns, autonomy spectrum, checklist, mistakes. EssayAgent EnablementThe operational layer — grounding, playbooks, the enablement stack, feedback loops. PillarLegacy SaaS RetrofitsConverting copilot-era products into agent-first ones without a rewrite. GlossaryTask DecompositionBreaking a goal into steps an agent can plan, execute, and check. GlossaryOrchestration vs AutomationWhen to script the steps, and when to let the agent decide them.
04
Technology · The Stack

What stack serves the work?

Technology comes fourth on purpose: the stack should serve the process design, not the other way around. Models are commodity — the decisions that matter are orchestration, tools, memory, runtime, and observability, and whether you build, buy, or orchestrate each layer.

PracticeAgentic API ExperienceThe API layer designed for the agent first — AX, the new DX. EssayAgent-First API DesignSix principles for designing APIs autonomous agents can use safely. ReferenceManaged Agent Platforms16 hosted agent platforms benchmarked — the build/buy/orchestrate landscape. EssayThe Two-Layer StackHow enforcement (governance toolkits) and design (AUX) compose in the stack. GlossaryModel Context Protocol (MCP)The connective tissue that lets agents reach systems without bespoke glue. GlossaryTool-Calling ArchitectureHow an agent acts on the world — and how to keep those actions safe. GlossaryMemory Systems for AgentsShort-term context, long-term memory, and the governance around both. GlossaryObservability for AI AgentsIf you can't see what the agent did, you can't trust, debug, or improve it.
05
AUX · The Trust Layer

What earns the trust that drives adoption?

This is the layer that wraps the other four and runs continuously. A safe agent isn't a trusted one — trust is a property of the relationship between your people and the agent, and it's earned, measured, and iterated. This is the half that compounds and can't be copied.

FrameworkThe 10 AUX HeuristicsThe agentic answer to Nielsen Norman — score the relationship, not the screen. FrameworkAgent-First Design PatternsThe complete catalog of 24 patterns across trust, control, memory, orchestration. KitTrustKit & The Trust HarnessThe discipline, the open-standard spec, and the runtime that enforces trust live. EssayA Safe Agent Isn't a Trusted OneMachine safety is commoditizing; the relationship is the moat. Why AUX is the edge. EssayDesigning Agentic User ExperiencesThe working method behind AUX — patterns instantiated as shipped product. EngagementAgent Experience AuditA structured diagnostic of a shipped AI product against the AUX framework. GlossaryExplainability in AgentsShowing the work — sources, uncertainty, logic — without overwhelming. GlossaryHallucination-Mitigation UXDesigning for graceful uncertainty: ask, escalate, return partial — never invent.

Five layers, moving together.

Skip one and the rest carry weight they can't hold. Wherever your gap is — still defining the vision, or scaling a live agent — we help you build the whole stack, with trust as the through-line.

Start a conversation → Read the 5-Step Playbook