Levi’s superagent: Orchestrating Employee Experience with Microsoft Foundry

11 6 minutes read

Levi Strauss & Co. (LS&Co.), the global apparel company known for its denim products, is developing an AI “superagent” as part of a broader digital transformation initiative. This superagent is designed to act as a centralized, conversational AI interface that automates and streamlines employee workflows across the organization, ultimately supporting LS&Co.’s goal of becoming a direct-to-consumer (DTC)-first, fan-obsessed retailer. Announced on November 17, 2025, the project emphasizes agentic AI—systems where AI agents can autonomously perform tasks and collaborate—to make complex operations more efficient and accessible.

Partnership and Key Technologies

LS&Co. is collaborating closely with Microsoft to build the superagent, leveraging Microsoft’s cloud and AI ecosystem for scalability and security. The core technologies include:

Microsoft Azure: Powers the backend infrastructure, including data migration from on-premises systems using tools like Azure Migrate. This enables a unified data backbone across over 400 technology systems, from sales floors to executive suites.
Microsoft Teams Integration: The superagent is embedded directly within Teams as an Azure-native orchestrator, allowing employees to interact via natural language queries in a familiar chat interface.
Microsoft 365 Copilot and Copilot Studio: Used for building custom AI agents, including rapid prototyping (e.g., one executive built a prototype agent called “Minerva” in an afternoon using Copilot Studio to analyze customer satisfaction reports).
Microsoft Foundry and Agentic AI Orchestration: Facilitates the coordination of multiple AI agents, enabling intelligent routing of tasks.
Semantic Kernel: Handles advanced automation, such as security agents and policy orchestration, while maintaining a zero-trust security model to protect data during AI operations.
Additional Tools: GitHub Copilot for developer productivity in areas like observability and quality engineering, and Microsoft Surface Copilot+ PCs with Windows 11 for secure device management via Microsoft Intune.

This tech stack allows the superagent to unify disparate systems, reducing reliance on legacy tools and enabling real-time insights.

Architecture: Superagent and Sub-Agents

At its heart, the superagent functions as an “intelligent intermediary” or orchestrator that routes employee requests to a network of specialized sub-agents. Rather than employees navigating multiple tools, they interact with one conversational portal in Teams, where the superagent intelligently delegates tasks behind the scenes.

Sub-Agents: These are modular, AI-powered components already deployed or in development across key functions. Examples include:
- IT support agents for troubleshooting and system queries.
- Human resources agents for employee onboarding, policy lookups, and insights.
- Operations agents for retail, warehouse, and supply chain tasks (e.g., inventory management or scheduling).
- Emerging agents for security and data governance.

The sub-agents collaborate seamlessly, drawing on shared data to provide context-aware responses. For instance, a query about store inventory might pull from retail sub-agents while cross-referencing HR data for staffing needs. This interconnected design ensures the superagent can handle repetitive, multi-step workflows—like generating reports or resolving cross-departmental issues—more efficiently than siloed tools.

Development Process

The creation process is iterative and phased:

Data Foundation: LS&Co. first migrated legacy on-premises data centers to Azure, consolidating environments and adopting zero-trust security to create a secure, scalable base for AI.
Agent Prototyping: Using Copilot Studio, teams rapidly build and test sub-agents. Some, like those for basic IT and HR tasks, are already live and handling real workloads.
Orchestration Build: The superagent layer is being developed to integrate these sub-agents into a single framework. Developers use GitHub Copilot to accelerate coding for integration and testing.
Testing and Iteration: Currently in build-and-test phase, with feedback loops from pilot users in corporate environments to refine routing logic and response accuracy.
Expansion: Beyond internal tools, the platform ties into customer-facing AI enhancements, such as personalized e-commerce recommendations and in-store experiences powered by similar agentic systems.

This approach emphasizes employee involvement, with AI designed to augment creativity rather than replace it—e.g., freeing designers to focus on innovation by automating routine data tasks.

Timeline

Current Status: Building and internal testing as of late 2025, with select sub-agents operational.
Initial Rollout: Early 2026 for corporate employees.
Global Expansion: Phased rollout to retail, warehouse, and international offices throughout 2026, aligning with LS&Co.’s $10 billion DTC revenue target.

Goals and Expected Impact

The superagent aims to “rewire” LS&Co. for a faster, more responsive future:

Employee Productivity: Reduce time on repetitive tasks by 20-30% (based on early Copilot pilots), enabling quicker access to insights and support.
Operational Efficiency: Streamline cross-functional workflows, from supply chain to retail, to support DTC growth and global fan engagement.
Customer Experience: Indirectly enhance consumer interactions by empowering store associates and e-commerce teams with real-time data, creating “connected, memorable experiences.”
Innovation Edge: Position LS&Co. as a tech-forward heritage brand, blending its 175-year legacy with AI to drive revenue and retention.

By centralizing AI in this way, Levi Strauss is betting on agentic systems to transform not just internal operations but the entire retail ecosystem, making every touchpoint—from design to delivery—smarter and more personalized.

Levi Strauss & Co. is accelerating its AI strategy through a partnership with Microsoft to build an AI orchestration platform.

This system includes a central “super-agent” supported by specialized sub-agents for departments like IT, HR, and operations, aimed at automating workflows, providing insights, and boosting efficiency.

The platform is scheduled for global rollout in early 2026.

Key AI Services:

For Employees: The orchestration platform will serve as an intelligent intermediary, streamlining support and automation across offices worldwide.
For Customers: The Levi’s app’s “Outfitting” feature, currently live in the US, Canada, and major European markets, delivers personalized styling recommendations. Upcoming enhancements in 2026 will include event-specific outfit suggestions.
For Retail Partners: The “Stitch” AI assistant, already in use via a mobile app in 60 US stores, offers quick access to product details, operational guidance, and training resources. A wider international expansion is planned for 2026.

Chief Digital and Technology Officer Jason Gowans emphasized that these AI tools are designed to make the business more agile, efficient, and innovative in the retail sector.

Stitch

STITCH is an AI-powered assistant developed by Levi Strauss & Co. specifically for retail store teams. Launched in late 2025, it aims to empower store associates and stylists by providing instant, conversational access to essential information, ensuring they can confidently assist customers without ever needing to say “I don’t know.”

The tool is part of Levi’s broader AI strategy to enhance operational efficiency and customer experiences across its global network of over 3,200 stores.

Key Features and Functionality

STITCH operates via an easy-to-use mobile app, leveraging a robust database that includes:

Product Information: Details on Levi’s offerings, such as fit comparisons (e.g., “How do our fits compare?”), differences between collections (e.g., “What makes the new Blue Tab™ collection different from the Red Tab line?”), and specifics like why selvedge denim stands out.
Operational Procedures: Step-by-step guidance on tasks like processing returns (e.g., “How do I process a return without a receipt?”) or enrolling customers in the Red Tab Loyalty program.
Training Materials: On-demand resources to upskill staff, covering everything from customer service best practices to merchandising guidelines.

Users can ask natural-language questions in real-time, receiving tailored responses that enable personalized advice, faster task completion, and more time for direct customer interactions. This “conversational” interface mimics a reliable teammate, supporting both new hires and seasoned employees alike.

Development and Technology

Origin: Conceived during an internal hackathon by Michael Buchanan, a 20-year Levi’s veteran who transitioned from store operations to the Data and Analytics team after completing the company’s in-house Machine Learning Bootcamp. Buchanan’s idea focused on bridging the gap between frontline needs and technology.
Partnership: Built in collaboration with Google Cloud, transforming the hackathon prototype into a fully functional app.
Responsible AI Focus: Developed with emphasis on data security, privacy standards, and risk mitigation to ensure ethical deployment.

Deployment and Rollout

Pilot Phase: Tested in 10 U.S. stores in early 2025, where it demonstrated quick wins in reducing response times and boosting staff confidence.
Current Status (as of December 2025): Rolled out to 60 U.S. stores ahead of the holiday season, with expansion to 50 additional U.S. locations underway as part of Levi’s 2025 growth initiatives.
Future Plans: Broader international rollout scheduled for 2026, aligning with Levi’s $10 billion revenue ambition. It will integrate further with other AI tools like the employee “super-agent” platform (in partnership with Microsoft) and consumer-facing features like the “Outfitting” app.

Impact and Feedback

Store managers have praised STITCH for accelerating onboarding and operational agility. For instance, Angel Mendoza, a store manager, noted: “Training takes time, and having instant access to accurate information means my team can confidently help shoppers even on their first day. STITCH is like a reliable teammate: ready to provide answers and support in real-time.” Early pilots showed improved customer satisfaction through more informed interactions, contributing to Levi’s goal of making retail experiences more seamless and innovative.

For more in-depth reading, check Levi’s official Unzipped blog post on STITCH.

azurecloud 3 hours ago