The Structural Shift in Experience Architecture
For the past twenty-five years or so, the fundamental logic of digital product design has operated on a singular, unidirectional axis: brands design an interface, and users adapt to it. Within this paradigm, organisations construct proprietary applications, websites, and walled gardens to capture attention, gate information, and drive transaction funnels. The framing interface around content—the chrome (browser tabs, OS patterns, viewport boundaries, and system gestures)—has been owned by platforms or platforms-as-brands to enforce corporate attention rules.
Under the guise of "user-centred design," UX teams have frequently functioned as an internal optimisation engine. Rather than purely serving the human, design has historically operated as a form of internal SEO for screens: interpreting messy user needs inside business constraints, translating them into flows, and optimising for business proxy metrics like clicks, conversion, and Net Promoter Scores (NPS1).
The slogan "putting the customer first" has often been rhetorical about the human, but structurally about the till.
This is the core of The Inversion of the Interface2. When individuals experience the world primarily through Agentic Mediation—deploying their own personal, foundational artificial intelligence layers—the primary UX layer is no longer a corporate application’s proprietary chrome. Instead, it shifts to the user’s delegated personal infrastructure.
A design leadership thesis for the agentic era
The strategic shift we are going through should be seen through the lens of an inversion of the interface: when users operate through a personal/foundational agent, the primary UX layer is no longer your app's chrome, it is the user’s delegated infrastructure.
This inversion shifts the purpose and meaning of design.
Design is no longer about surface optimisation wrapped in empathy language.
The question is no longer “did the flow feel smooth?” as instead: “was authority clear, bounded, recoverable, contestable, and governable?”
If that infrastructure (a clear, bounded, recoverable, contestable, governable authority) fails, persuasion won't save you, only operational truth can save you.
Designers move from persuasion to constitutional work
In an agent-mediated world:
Authority becomes the primary UX object
Design decides what can be delegated, by whom, under which constraints, and with what reversal rights.
(Concepts: Consent and control, Progressive autonomy, Authority Design, Governable Delegation)
Meaning becomes infrastructure
If value cannot be represented in machine-legible semantics, your brand cannot compete except through residual friction and lock-in.
(Concepts: State model, Operational truth, Common core + local layers)
Failure becomes first-order experience
Agent systems are judged by interruption safety, reversibility, and escalation quality more than by visual polish.
(Concepts: Recoverability, Escalation paths, No Silent Failure)
Legitimacy becomes product scope
If users cannot challenge outcomes and win when right, you do not have trust; you have deferred risk.
(Concepts: Evidence surfaces, Auditability, UX, Contestability)
Governance becomes delivery acceleration
Without decision rights, quality gates, and decision artefacts, teams drift into prompt theatre and late reversals.
(Concepts: Governance that ships, Decision Artefacts, Decision Rights, Quality Gates)
Governable Delegation Framework (GDF-5)
Use this as the schematic for “what designers will do” in AI-dominant organisations.
| Domain | Core Question | Foundational Concepts | Representative Patterns | Primary Artefacts |
|---|---|---|---|---|
| Authority Design | What may the agent do, under whose authority, and within which limits? | Consent and Control · Progressive Autonomy · Trust Cues | Autonomy Gradient · Guardrails UX · Delegation Models | Delegation Contracts · Autonomy Ladders · Permissions Matrices · Refusal Patterns |
| Semantic Design | How is meaning preserved across systems, organisations and agents? | Operational Truth · State Models · Common Core + Local Layers · Agent Experience (AX) | State Models for Agent Workflows · Semantic Contracts | State Machines · Entity/Event Schemas · Policy Objects · Protocol Specifications |
| Recovery Design | What happens when reality diverges from expectation? | Recoverability · Escalation Paths | Handoff & Escalation Design · No Silent Failure | Failure Taxonomies · Retry & Rollback Logic · Escalation Packages · Operational Runbooks |
| Legitimacy Design | Can outcomes be understood, challenged and corrected? | Evidence Surfaces · Auditability UX · Contestability | Evidence Surfaces · Contestability Flows · Audit Trail UX | Evidence Views · Appeal Flows · Audit Interfaces · Correction Loops |
| Evolution Design | How does the system adapt without sacrificing coherence or control? | Governance that Ships · Decision Artefacts · Quality Gates | Adoption Operating Rhythm · Governance by Design | Gate Checklists · Decision Logs · Operating Cadences · Policy & Release Notes |
Design mandate in one line:
From shaping screens to shaping the constitutional, semantic, and operational conditions under which delegated action remains legitimate.
Role architecture: what design teams become
In GDF-5 terms, each role owns a part of delegated authority.
| Role | Primary Ownership | Non-Negotiable Outputs | Failure Mode if Missing |
|---|---|---|---|
| Head of Agentic Design | System coherence | Autonomy policy, decision rights, governance model | Local optimisation, fragmented initiatives, no coherent operating model |
| Policy & Promise Designer | Delegation boundaries | Executable policies, exception paths, dispute mechanisms | Service behaviour drifts from stated promises |
| Semantics & Data Contract Designer | Shared meaning | Schemas, event models, interoperability contracts | Inconsistent interpretation across agents, systems and organisations |
| State & Recoverability Designer | Operational continuity | State models, recovery logic, rollback mechanisms | Irrecoverable errors, duplicated actions, hidden failure states |
| Trust, Consent & Contestability Designer | Legitimate delegation | Permission models, evidence surfaces, challenge and appeal flows | Trust erodes when outcomes cannot be questioned or corrected |
| Outcome Experience Designer | Outcome legibility | Receipts, status visibility, reason codes, confidence signals | Users experience systems as opaque regardless of technical quality |
| Governance & Quality Gates Lead | Sustainable delivery | Quality gates, governance cadence, decision artefacts | Continuous rework, governance debt, operational instability |
A practical maturity model for design leaders
Use this to position teams and plan capability build.
| Level | Maturity Stage | Description |
|---|---|---|
| 0 | Decorative AI | AI is primarily used as a surface-layer enhancement. No explicit autonomy model exists, and explainability is treated largely as a communication exercise. |
| 1 | Assistive AX | Systems provide suggestions, drafting support, and task assistance. Guardrails remain largely static, with limited visibility into state, reasoning, or delegation. |
| 2 | Controlled Delegation | Specific tasks can be delegated under defined conditions. Approval checkpoints, recovery mechanisms, and human oversight are built into the workflow. |
| 3 | Governable Autonomy | Autonomous actions operate within a framework of contestability, auditability, escalation ownership, and formal quality governance. |
| 4 | Protocol-Native Organisation | Services, policies, and brand value are represented through machine-readable semantics that allow trusted negotiation and execution across independent agent ecosystems. |
Promotion Gate Criterion:
Advancement between levels must be blocked unless specific Agent Experience (AX) acceptance criteria are met, explicitly derived from: Autonomy Gradient, State Model for Agent Workflows, No Silent Failure, Contestability Flows, Audit Trail UX, and Adoption Operating Rhythm.
What changes for the profession itself
Small-d design does not disappear, it industrialises
UI grammars are stabilising into systems of automated stewardship, while valuable execution work remains, less of it defines a firm’s long-term strategic advantage.
Capital-D Design expands
The high-leverage frontier shifts completely to systemic, structural engineering:
- institutional interface design (policy frameworks, responsibility matrices, structural recourse)
- semantic protocol design (state models, behavioral constraints, transaction guarantees)
- governance design (quality gates, structural artifacts, operating rhythms)
The new Designer is a legitimacy engineer
Not in the technical sense only, but in the civic/product sense:
they design how power is delegated, inspected, contested, and corrected.
KPI stack
Delegation integrity: % of agent actions executed with an explicit scope, bounded autonomy level, and pre-defined rollback class.
Recoverability performance: Target time-to-recovery, system reversal latency, and self-serve user recovery rates.
Legitimacy throughput: Mean time to dispute resolution, reversal quality score, and correction loop completion rates.
Audit usability: Time required for an internal or external auditor to reconstruct an agent’s decision path; % of investigations completed successfully without engineering intervention.
Governance lead time: Velocity from a policy or UX loophole discovery to a controlled, audited system release.
No-silent-failure index: Proportion of structural failures discovered and isolated by system logs/operations before causing user-window harm.
Operating rhythm
The minimum viable governance is about shipping: short cycles, explicit owners, and rigorous artefact discipline.
Weekly: Reliability + Exceptions Board – Cross-Functional Attendants: Design, Engineering, Operations, Risk.
Outputs: identification of failure clusters, stuck states, edge-case escalations, and recovery deltas.
Monthly: Policy/Semantics/State Board – Governance: Chaired explicitly by Design.
Outputs: versioned policy objects, state map diffs, data contract changes, and quality gate updates.
Quarterly: Autonomy Promotion Review
Outputs: Promotion or demotion of automated decision classes, empirical evidence packages for autonomy expansion, and active rollback drills.
Leadership position
If agents are to become the dominant mediation layer of human activity, design cannot remain confined to the optimisation of attention surfaces: its existential mandate is to define the terms of delegated action.
The future of designers is about authoring, structuring, and operating the behavioural frameworks that make autonomous actions legible (state + evidence), bounded (permissions + guardrails), recoverable (rollback + escalation), contestable (appeal + correction), and operable (governance + quality gates).
That is the practical path for the discipline to move “for the user” away from empty corporate rhetoric and turn it into permanent technical architecture.
✺✺✺
References
Inverting the interface
What happens when people experience brands, services, and institutions primarily through their own AI agents—rather than through brand‑designed apps, websites, and campaigns?
Age of mediocrity: designers and the AI mirror
A reflection on the future of Design and the evolving role of the Designer in a world increasingly shaped by artificial intelligence.
✺
Disclaimer
This post was not generated by AI.
Artificial intelligence was used selectively during the research phase, primarily for exploratory tasks, and sparingly during the writing phase to review and refine English fluency. When used in this post, AI is explicitely marked.
All data from publicly available sources.
All opinions expressed in this article are solely my own and do not represent the views of any current or former employer.