How does Tapestry use data and AI for demand planning and personalisation at Coach?
Tapestry, owner of Coach and Kate Spade, built a shared data layer, used AI to forecast demand and position inventory, and now runs an in-house generative AI system called Mira for planners and merchants.
KEY TAKEAWAYS Summary by the editors
- Tapestry CEO Joanne Crevoiserat said on the company's fiscal Q2 2023 earnings call that it was using artificial intelligence to forecast customer demand, which led to higher inventory availability.
- Tapestry kept Coach's Tabby bag rather than marking it down after its data showed younger consumers gravitating to it, Crevoiserat told Fortune in November 2024.
- Mira is Tapestry's in-house generative AI system for employees, built on Amazon Bedrock, which surfaces insights for assortment planning, inventory and trend tracking in seconds to minutes instead of days.
- Mira runs on Tapestry's patented Global Data Fabric with role-based access, and Tapestry has received a US patent for Mira's core architecture, according to Digital Commerce 360.
- Tapestry's digital product creation team uses 3D digital twins of products such as the Tabby, which the company says cut one project from two years at a supplier to two weeks to design approval.
Tapestry uses AI in two layers at Coach and Kate Spade: machine learning to forecast demand and position inventory, and an in-house generative AI system, Mira, that lets planners and merchants query company data in plain language. Both rest on a shared data platform and direct-to-consumer customer data, which is why the company's executives describe AI as the result of years of data investment rather than a separate project.
How does Tapestry use AI to forecast demand?
Tapestry has described AI forecasting since at least its fiscal 2023. According to Consumer Goods Technology, CEO Joanne Crevoiserat said on the Q2 2023 earnings call that the company was "utilizing artificial intelligence to forecast customer demand" and to position inventory and stores, and that "this led to an increase of inventory availability". The same report noted that this was the first holiday season with all of Tapestry's brands on its digital enterprise platform, which supplied the consistent data the models needed.
Coach CEO and brand president Todd Kahn said data informs Coach's pricing, promotions and placement: "It really informs everything we do." For merchandisers, the practical value of demand forecasting at this level is allocation: which store or channel gets which units, and when, so that full-price sell-through improves without excess stock. Tapestry has not published model accuracy figures, so the public evidence is the outcome executives described (better availability) rather than a measured forecasting benchmark.
How did data change decisions on the Coach Tabby?
The Tabby bag is Tapestry's most cited example of data informing a product decision. In a November 2024 interview with Fortune, Crevoiserat explained that sales of the bag were slipping during the pandemic, but data showed that "this is a bag that younger consumers are gravitating to", so the company kept it instead of marking it down. A softer Pillow Tabby followed in 2021. Crevoiserat described the approach as "the intersection of data that informs the creativity".
This is not an AI story in the narrow sense, and Fortune did not describe it as one. It is a merchandising story about reading customer data correctly, which is the decision AI forecasting tools are supposed to support.
What is Tapestry's Mira AI system?
Mira is an internal generative AI system built by Tapestry for employees, not shoppers, according to Digital Commerce 360 in May 2026. It pulls data from across the business and surfaces insights "in seconds to minutes"; previously, staff gathered information manually from several dashboards, which could take days. Employees use it to plan assortments, manage inventory and track consumer trends.
Several design choices stand out:
- Data foundation: Mira works on top of Tapestry's patented Global Data Fabric, which connects data sources across the business.
- Domain knowledge: it is trained on retail and fashion terminology and institutional knowledge.
- Access control: it runs in a secure enterprise environment with role-based access, so employees see only the data they are authorised to view.
- Infrastructure: deployment uses Amazon Bedrock, a managed service for building generative AI applications.
- Intellectual property: Tapestry received a US patent for Mira's core architecture, its first AI patent.
Chief data and analytics officer Fabio Luzzi called Mira "the result of years of deliberate investment in our data infrastructure". Crevoiserat described AI as "a competitive advantage" on the company's Q2 fiscal 2026 earnings call, with AI tools embedded across the value chain from product development to pricing and marketing.
| Use | What it does | Reported by |
|---|---|---|
| Demand forecasting | AI forecasts customer demand to position inventory and stores | Consumer Goods Technology (2023) |
| Customer insight | Data on how customers enter the brands and what they buy next | Consumer Goods Technology (2023) |
| Mira | Internal generative AI for assortment, inventory and trend questions | Digital Commerce 360 (2026) |
| Gift Assistant | Conversational AI on KateSpade.com mobile site to help find gifts | Digital Commerce 360 (2026) |
| 3D digital twins | Digital product creation for design, sampling and marketing assets | Adobe (2025) |
How does Tapestry personalise for customers?
Personalisation at Tapestry starts with identification. In 2023 Crevoiserat said the company understands "how customers are coming into our brands, what they're likely to purchase next", which helps drive purchase frequency. Digital Commerce 360 reports that Tapestry's 2025 annual report lists customer journey personalisation among its AI focus areas and says the tools improve how it identifies and segments customers. The most visible customer-facing tool is Kate Spade's Gift Assistant, a conversational AI on the KateSpade.com mobile site. Luzzi said it came "from listening to our consumers and figuring out what they actually needed".
Where do 3D and digital twins come in?
Tapestry's digital product creation team in New York, which merged with its 3D team in 2021, builds digital twins of products under a "create once, use many" approach, according to an October 2025 Adobe case study. The team reports that a project stalled at a supplier for two years moved from concept to design approval in two weeks, and that prototyping costs for the Jelly Tabby mould fell to a fraction of their earlier level. The same 3D assets feed marketing, retail and e-commerce imagery. Accurate digital products also improve the data that planning tools use.
What can merchandisers learn from Tapestry?
- Unify data across brands and channels first; Tapestry's forecasting and Mira both depend on a shared platform.
- Use AI to speed up questions planners already ask, such as assortment and inventory, rather than inventing new workflows.
- Build access control into internal AI from the start so sensitive data stays with authorised roles.
- Keep human judgement on product decisions; the Tabby case shows data informing creative choices, not replacing them.
Frequently asked questions
What is Mira at Tapestry?
Mira is Tapestry's in-house generative AI system for employees. It pulls data from across the business so staff can plan assortments, manage inventory and track trends, returning insights in seconds to minutes instead of days, according to Digital Commerce 360.
Does Coach use AI for demand forecasting?
Yes. Tapestry's CEO said in fiscal 2023 that the company used AI to forecast customer demand and position inventory, which increased inventory availability. Tapestry owns Coach and Kate Spade.
How did data help the Coach Tabby bag?
During the pandemic, Tabby sales slipped, but Tapestry's data showed younger consumers were drawn to it, so the company kept it instead of marking it down. A Pillow Tabby followed in 2021, according to Fortune.
Which AI platform does Tapestry use?
Mira is deployed on Amazon Bedrock, AWS's managed service for generative AI applications, and works with Tapestry's own Global Data Fabric, according to Digital Commerce 360.
One edition every weekday morning. Read in five minutes. Free for industry professionals.
