The sector enters this benchmark at the upper edge of Operational. LVMH's MaIA — reportedly serving over 40,000 employees at more than 2 million requests monthly — is the clearest evidence that advisor augmentation has moved from pilot to daily workflow at the pinnacle tier. This is the Decision Surfaces framework made concrete: AI sits beside the client advisor, surfacing histories and preferences rather than replacing the human. But access is concentrated. Outside the largest groups, adoption is described as bottom-up and informal, and privately held houses (Chanel, Hermès) disclose little. The sector level reflects real, scaled access at the top and thinner, uneven access below it.

The defining tension of the fashion sector's AI posture is captured in one pair of figures: 92% of fashion companies plan to increase AI investment, yet only 1% report mature deployment (Alhena). This is a sector rich in intent and thin in scaled execution. Under the Compiled Corporation lens, only the pinnacle groups — LVMH's centralised platform across 75 Maisons, Burberry's real-time forecasting — are genuinely compiling core decisions into systems. The rest remain in pilot purgatory. The sector sits just inside Operational, held there by a small number of leaders rather than broad-based maturity.

Governance is where the sector is weakest relative to its ambitions, sitting in Active. The pressure is largely external: New York's synthetic performer disclosure law (effective June 2026) and its posthumous right-of-publicity statute impose hard compliance obligations on AI-generated campaign content. Chambers frames robust governance as an emerging strategic pillar — telling language for a capability not yet built. The Gucci 'Primavera' backlash is the cautionary datum: AI creative deployed without adequate guardrails on the Identity Control Surface, where non-human-generated identity meets a craft-built brand promise. The sector is being governed by regulators before it governs itself.

Talent build-out is visibly underway, placing the sector in mid-Operational. AI roles are reported among the fastest-growing across all five named pinnacle groups, and Kering's dedicated venture arm points to structural, not opportunistic, investment. What the evidence does not yet show is systematic reskilling of the existing workforce — the artisans, merchandisers and store staff whose judgment the Compiled Corporation must eventually encode. Hiring AI specialists is easier than upskilling a craft workforce; the sector is doing the former faster than the latter. Level held at the boundary between building talent and transforming the broader workforce.

ROI evidence is stronger here than in most dimensions, holding the sector in upper Operational. The forecasting use case is the clearest win: Burberry's real-time demand forecasting delivers measurable sell-through gains, and sector-wide, forecast-error reductions of 20–50% translate directly to markdown savings. Kearney reports luxury executives now tie AI budgets to conversion, AOV and retention — a discipline that separates measured returns from AI-washing. The caveat: the largest figures (McKinsey's $150–275B) are potential, not realized, and vendor case studies (StyTrix's 747% ROI) warrant skepticism. The level reflects proven operational ROI at leaders, not yet enterprise-wide financial impact.

Fashion has arrived at a sophisticated, consistent AI narrative, placing it in Operational. The messaging is deliberately understated — Cucinelli's 'handmaiden,' LVMH's 'at the service of our houses' — a posture Kearney and FIU both characterize as AI's 'quiet revolution.' This is the Janus Brands discipline done well: AI messaging engineered to sit behind the craft narrative rather than compete with it. The risk is that quiet fluency shades into invisibility, ceding the AI-native conversation to challengers and platforms. But as a matter of consistency between AI messaging and legacy identity, the sector's leaders are executing with rare coherence.

This is the sector's clearest structural weakness, sitting in Active. The Identity Control Surface and agent-readiness converge here: AI shopping agents read structured attributes, not the aspirational prose luxury brands have perfected for humans. Mapp and Mirakl both document that fashion catalogues largely fail the machine-readability bar — and Adobe's data makes the cost concrete: AI-referred traffic grew 393% YoY and now converts 42% better, but only for merchants agents can parse. Beauty peers (Ulta, Glossier) are already live on commerce protocols; fashion is not. Luxury's craft-led aversion to standardized product data is precisely the trait that leaves it least legible to agents. The level is low deliberately: intent is high, infrastructure is not.

Market perception of fashion's AI competence sits at the Operational/Systemic boundary, propped by the pinnacle groups. LVMH and Richemont draw sustained, substantive third-party analysis — real deployments, named partnerships, measurable scale — that reads as competence rather than AI-washing. But the Gucci 'Primavera' episode is a warning under the Janus Brands lens: for a craft-built house, visible AI can subtract credibility as fast as it adds it. Perception is bifurcated — strong for back-office and clienteling AI, fragile for front-of-house creative AI. The level reflects earned credibility at the top, tempered by the sector's unresolved public ambivalence.

Fashion content is optimized for aspiration and human browsing — the opposite of what AI consumption requires — placing this dimension in Active, the sector's lowest. Merkle's finding is blunt: human- and Google-optimized copy does not work for agents. Luxury's editorial, mood-led product storytelling is the least agent-legible content format in commerce. Kearney's warning lands hardest here: in the agentic era, visibility is earned through 'clarity, data integrity and trust, not brand heat alone.' The dimension shares no direct routing signal in this cut, so it is scored on the abundant adjacent agent-commerce evidence — and that evidence points one direction. This is where the Decision Surfaces shift from human eyes to agent parsers, and fashion's content has not made the transition.

Brand and marketing operations are being assisted by AI, not yet orchestrated by it, placing the sector in Active. LVMH scales personalized marketing content with generative AI, and Kering applies AI to precision media buying — but these are point solutions, not autonomous brand operations. Glossy's finding of channel-by-channel hesitancy (fast in social/retail media, slow in influencer/CTV) confirms orchestration is fragmented. The human still holds the marketing Decision Surface across most of the funnel. The level reflects meaningful AI in the marketing stack without the enterprise brand-ops automation that defines the top of this scale.

The consumer-facing interaction layer is advancing from chatbot toward assistant, sitting in Operational. Zalando's conversational assistant and Kering's 'Madeline' show fashion — including luxury — building genuine assistant-grade Decision Surfaces where shoppers describe intent rather than navigate categories. Demand is real: over 40% of Gen Z and Alpha use AI weekly for fashion. But most deployments remain assistant-grade (respond to prompts) rather than agentic (act autonomously within a role) — Algolia's framing of the assistant-to-agent distinction is the gap. The level reflects credible assistant deployments with agentic capability still emerging.

Scaled agentic systems in production remain rare, placing the sector in Active. Named deployments exist — Kering's Madeline, Zalando's assistant — but these are early and largely assistant-grade; the 1% maturity figure applies with force here. The Compiled Corporation endpoint (autonomous agents completing transactions and decisions) is where Mirakl says commerce value is heading, and where fashion is not yet. Notably, 72% of CEOs intend to keep humans in the loop through 2030 — a deliberate governance choice on the Identity Control Surface that both slows and disciplines agentic scaling. The level reflects real pilots without production-scale autonomy.

AI UX in fashion is philosophically sophisticated but unevenly executed, sitting in mid-Active/Operational. The sector has articulated the right standard — FIU's 'invisibility,' LVMH's advisor augmentation that deepens rather than displaces human connection, Algolia's intent-understanding search. These are mature Decision Surface design principles. But the Gucci backlash shows trust and error recovery break down when AI surfaces to the consumer without adequate framing. The gap between UX philosophy and consistent delivery keeps this dimension below its brand-messaging counterpart. The level reflects strong intent and leader-tier execution against inconsistent front-of-house trust outcomes.

Process redesign is genuine at the leading edge, placing the sector just inside Operational. Kearney's evidence that AI is embedded across forecasting, design, clienteling and service — and Kering's redesigned store-to-store inventory and trend-prediction workflows — show the Compiled Corporation logic taking hold in core operations, not as bolt-on features. McKinsey's caution frames the ceiling: the real advantage goes to firms that redesign business models, not just workflows, and most have not. The forecasting and inventory functions are being rebuilt around AI; design and creative processes remain deliberately human-anchored. The level reflects real operational redesign at leaders against surface-level adoption elsewhere.

Measurable value from product AI is now documented, placing the sector in upper Operational. The Adobe data is the strongest single datum in this cut: AI-referred traffic up 393% YoY and converting 42% better — a reversal from a year prior. Personalization's 10–15% revenue lift is a realistic, measured return, and Crocus's 86% deflection with 84% CSAT shows focused deployments delivering. The consistent theme: value accrues to brands whose product data and interaction layers are ready to capture agent-driven demand — tying this dimension directly to the sector's Agent-Ready Infrastructure weakness. Value is real and measurable at the prepared; the level reflects that value is being delivered, not uniformly captured.