We are turning this engagement into a practitioner playbook: the orchestration architecture, the enterprise-context design, the disagreement and governance patterns, and the economics scaffolding that turns a market signal into an investment-grade decision.
Coming soonA Fortune 500 global apparel and lifestyle company wanted to know whether agentic AI could compete with its traditional new-product innovation process.
The company's standard process took approximately 18 months. A serious new-product opportunity would require consumer insight, category and competitive research, product and industrial design, materials, sourcing, manufacturing planning, pricing, channel strategy, marketing, ecommerce, unit economics, and ultimately a complete P&L.
Much of that expertise already existed inside the organization. The problem was getting it to operate as one system.
As a largely sequential process, a question from one function generated another round of analysis somewhere else. Information risked falling through the cracks between teams. Different functions sometimes worked from different assumptions. Decisions remained largely siloed, leaving the Innovation Office to reconcile inputs across 15 to 20 teams.
Annually, the Innovation Office typically developed two or three concepts deep enough to recommend one, while the broader enterprise might make a major investment behind a new concept roughly once every two to three years.
For a global retail company, the constraint matters in a specific way. A concept is not valuable simply because consumers like it. It must also survive questions about brand fit, materials, cost, sourcing, margin, channel, timing, inventory risk, demand, and whether the economics justify putting the product into the line.
The constraint was not idea generation. It was the organizational capacity to research, challenge, economically underwrite, and align around an idea well enough to commit scarce people, sourcing capacity, manufacturing capacity, and capital.
This is consistent with broader pressure across fashion and retail. McKinsey's State of Fashion 2026 argues that traditional advantages such as scale and low-cost sourcing are no longer sufficient, and that fashion companies increasingly need AI and new technology to improve productivity and reinvest in differentiated growth. Deloitte's 2026 merchandising research similarly finds that organizational complexity and misaligned priorities are preventing many retailers from fully translating data and AI into faster, better product and assortment decisions.12
The VP of Innovation, working within the COO organization, was not looking for another AI demonstration. Leadership wanted to understand what AI could actually change about the product innovation system.
The mandate centered on specific questions. Could AI identify attractive market and consumer opportunities more continuously? Could it develop credible product concepts grounded in real consumer, category, and cultural signals? Could those concepts be evaluated against the company's proprietary product, brand, sourcing, manufacturing, marketing, and financial information? Could AI materially accelerate the work required before leadership committed designers, product teams, sourcing teams, factories, marketing resources, and capital?
The benchmark was not another AI tool. It was the company's existing way of developing products.
A conventional approach would have introduced AI function by function. Design gets an AI tool. Consumer insights gets another. Marketing gets a copilot. Merchandising experiments separately. Supply chain builds its own models. Finance builds another. Everyone gets somewhat faster while the same organizational architecture remains intact.
We recommended a different approach.
Centralize orchestration. Distribute expertise.
Functional teams would retain their domain expertise and decision authority. The COO Innovation Office would retain the orchestration layer. Specialized agents would perform increasing amounts of the research, analysis, synthesis, scenario development, and coordination required between functions.
The resulting architecture sat between two common extremes. It was not a centralized AI organization attempting to replace product, design, marketing, merchandising, or supply-chain teams. And it was not dozens of employees independently prompting copilots. It was an agentic product-innovation system designed to connect specialized expertise around a common set of investment decisions.
The industry evidence points in the same direction. Deloitte's 2026 survey of 200 retail and consumer executives found that 75% call AI a top strategic priority, but only 16.5% can quantify a return, while enterprise-wide deployment remains in the 7 to 10% range. Deloitte's conclusion is blunt: strategy is moving faster than operating models, and pilots are moving faster than production.3
A seven-person multidisciplinary team spanning AI architecture, product innovation, consumer intelligence, creative, growth, customer experience, and product leadership designed the operating architecture.
The resulting system used more than 40 specialized agents across the major stages of product innovation. Their capabilities covered market and trend intelligence; consumer segmentation and validation; competitive and category analysis; whitespace identification; product concepts and design; materials and construction; sourcing and manufacturing planning; bills of materials and technical specifications; compliance; channel and direct-to-consumer strategy; media and influencer planning; pricing; demand forecasting; customer-acquisition economics; conversion scenarios; margin optimization; three-year P&L scenarios; and portfolio investment recommendations.
The number of agents was not the important part. The architecture connecting them was.
Fashion leaders are beginning to describe the same shift. In June 2026, ASOS's CTO called AI scaling primarily an organizational challenge, not a technology challenge, and said the company is extending agentic AI into buying, design, and merchandising. His warning is directly relevant here: low-value agents proliferate easily unless leadership redesigns the operating model around commercially meaningful use cases.4
The system was designed to work with information normally scattered throughout a global apparel company.
Customer segmentation, purchase behavior, reviews, focus groups, social listening, search behavior, social-platform signals, and cross-brand customer overlap.
Historical product catalogs, SKU performance, sell-through, markdown history, category trends, competitive pricing, whitespace analysis, seasonal line plans, and prior product decisions.
Brand positioning, design language, seasonal direction, photography standards, naming conventions, and historical design-review decisions.
Approved materials, component specifications, supplier information, factory capabilities, capacity, quality metrics, landed costs, historical tech packs, and bills of materials.
Unit economics, target margins, product-line P&Ls, investment thresholds, and historical revenue performance.
Generic AI could generate a bag concept. That was not the objective.
The objective was to allow AI to reason about a product opportunity using information a competitor's model would not possess: this company's consumers, brands, products, materials, suppliers, factories, margins, historical launches, and market performance. The foundation model was broadly available. The enterprise context was not.
The agents did not operate as independent chatbots or AI tools. Outputs from one part of the system became structured inputs into another.
A new-product investment recommendation might require: consumer need, then product concept, then brand fit assessment, then pricing, then material and manufacturing cost, then channel strategy, then demand assumptions, then customer-acquisition economics, then margin analysis, then scale economics, then a three-year P&L.
Inputs came from different parts of the organization. Product defined the use case. Consumer insights validated the need. Design shaped the concept. Supply chain supplied material and manufacturing economics. Growth supplied media assumptions. Ecommerce contributed conversion and returns assumptions. Financial agents reconciled those inputs into scenarios and an investment recommendation.
When a function surfaced new evidence contradicting an earlier conclusion, the system did not simply average the disagreement away. The evidence could be pushed back through the workflow. Relevant agents reconsidered the original assumptions, recalculated the economics, and surfaced the trade-offs again.
Humans owned consequential decisions and were accountable for them. But substantially more of the work required to reach those decisions could happen continuously.
Perhaps the most revealing test was not whether the agents could produce an attractive product but also whether they could recommend against one.
The system developed multiple AI-generated product opportunities and compared them with a product concept developed through the company's conventional innovation process. The portfolio analysis did not simply rank the AI-generated ideas first.
One AI-generated concept was recommended as the lower-risk market-entry opportunity. The conventionally developed concept was judged to have stronger long-term revenue potential. Another AI-generated concept was held back because important customer-acquisition, conversion, and economic assumptions had not yet been sufficiently underwritten.
The system was not being rewarded for producing more products. It was being asked: which concepts deserve a place in the portfolio, and which deserve investment?
Leadership could compare opportunities using a common view of demand, customer economics, gross margin, risk, readiness, and three-year P&L instead of waiting for separate functional analyses to be reconciled afterward.
The clearest result was not an AI-generated product. It was a different way of making product-innovation decisions.
The AI-generated concepts were well received by leadership, particularly because the recommendations could be connected back to real market signals and enterprise-specific information. Leadership accepted the strategic capability and chose to continue operationalizing the agentic approach as part of the innovation workflow.
Leadership estimated that approximately 12 months of planning, analysis, follow-up questions, and cross-functional coordination could potentially be compressed into weeks.
This was not a claim that an 18-month physical product-development cycle suddenly became a few weeks. It was a more specific and strategically more important claim: much of the organizational work required to determine which concepts deserved design, sourcing, factory capacity, inventory investment, marketing support, and capital could happen dramatically faster.
That has particular significance in apparel and fashion. The economic cost of a poor product decision does not end with development expense. It can show up as inventory, markdowns, missed seasonal windows, working capital, wholesale commitments, lost shelf space, and brand dilution. The ability to kill weak concepts earlier can therefore be as valuable as the ability to launch strong concepts faster.
The experiment also clarified where AI stopped creating the same degree of leverage.
Agents could contribute to product specifications, bills of materials, tech packs, compliance analysis, sourcing options, and preparation for prototyping and sampling. But the physical product system still contained constraints that intelligence alone could not eliminate: factory scheduling, factory capacity, tooling, material sourcing, physical samples, testing, and quality assurance.
AI had accelerated the information economy of product innovation. The physical product still had its own clock.
For apparel and fashion leaders, that creates an important strategic distinction. AI may allow the organization to decide what should be made dramatically faster. That does not mean the supply chain can immediately make it at the same speed.
A second limitation emerged only after the system became faster. Human decision gates began arriving much more quickly. A function that previously had days or weeks between rounds of analysis could receive an updated recommendation, revised economics, or a new trade-off far sooner.
AI increased the supply of decision-ready analysis. It did not automatically increase the organization's capacity to absorb and act on it.
Governance therefore became more consequential. Leadership needed to understand which data changed, which assumptions changed, why a recommendation changed, which functional agents contributed, and where human judgment entered the process. The audit trail was not administrative overhead. It became part of the innovation operating model.
The easier path is to let every function have its own AI tools. Design gets faster. Marketing gets faster. Consumer insights get faster. Supply chain gets faster. Finance gets faster.
Yet the underlying problem remains. The same information sits in separate functions. The same handoffs exist. The Innovation Office still reconciles conflicting assumptions. And the portfolio decision still depends on assembling the entire picture afterward.
Individual productivity does not automatically create organizational throughput. The unit of transformation was not the employee. It was the product-innovation system.
The company began with a technology question: how much can agentic AI actually do? It ended with an organizational question: how should product innovation work when analysis and coordination that once required months can happen in weeks?
The emerging model was straightforward. Functions own the expertise. Humans own consequential decisions. Agents perform and challenge the analysis. The Innovation Office orchestrates the system.
The strategic advantage was not 40 agents. It was lowering the organizational cost of getting from a consumer or market signal to an investment-grade product decision.
For an apparel and fashion company, that changes more than cycle time. It changes how many consumer opportunities the company can afford to investigate before committing design resources, supplier capacity, inventory, marketing spend, and capital.
We are packaging the orchestration architecture, the enterprise-context design, and the disagreement and governance patterns from this engagement into a practitioner playbook. Coming soon.
Widespread AI adoption without materially changing innovation throughput. Employees would become faster. Functions would generate more analysis. More ideas might surface. But the same organizational machinery would still determine how quickly a consumer signal could become a funded product opportunity.
Functions keep expertise and authority. A 40+ agent layer researches, challenges, and economically underwrites concepts against proprietary data. Leadership compares opportunities on one shared view of demand, economics, risk, and P&L.
The risk was not failing to adopt AI. It was putting AI inside a product-development process designed before AI existed.
The unit of AI transformation is not the employee — it is the system the work moves through. Redesign how consumer insight, design, merchandising, sourcing, manufacturing, marketing, and finance connect around a decision, ground the agents in your own data, and the cost of getting from a consumer signal to an investment-grade product bet drops sharply.
We help COOs, innovation leaders, and product leaders design AI operating models — the agent architecture, the enterprise-context layer, the decision rights, and the governance — then run them against real work.
Start a Conversation