My Favorite Quote of the Show:
Being able to move faster and being able to be more specific to a trend in the moment has become that much more important because consumer behavior and consumer trends change at the speed of light. So even trying to react to them in this day and age without some sort of AI and agentic tools is a fool's errand. - Sue McMahon, Global Strategy Director | Merchandising & Revenue Growth Microsoft.
What’s it all about? My Take
Why "Autonomy" has been a Dirty Word in Retail (And How AI Fixes It): Merchants and planners naturally resist black-box decision-making, as complete, unsupervised autonomy removes human accountability. AI tools can resolve this by providing governed intelligence with transparently surfaced data and clear guardrails. The strategic control is still in human hands.
The Death of Siloed Data & Microsoft's Platform Neutrality: Siloed data had historically forced companies into costly, multi-year migration projects. Platforms like Microsoft Fabric provide a lakehouse architecture that queries data where it sits neutrally whether across AWS, Snowflake, Databricks, or Google. This neutrality allows retailers to bypass data consolidation paralysis and, instead, deploy actionable AI directly on top of their existing baseline.
Building Proof and Trust for Technological Adoption: I learned in this discussion that technology can move faster than organizational trust.
Ask Ralph built developed in partnership with Microsoft and runs on the Microsoft Azure OpenAI platform - One Year Later
Ask Ralph, one year later: In the three post earnings calls since the launch of Ask Ralph in September 2025, CEO Patrice Jean Louis Louvet referred to its contribution of First Party Data as “gold” and did not attribute sales to the app. The app launched in Q2 FY26 and company revenue in total has increased: Q2 FY 26 sales were up 17%, Q3 FY 26 +12.2%, Q4 FY26, +17%. The total FY26 was up 15% to 8.1 billion in revenue.
The app was concepted by David Lauren, Chief Branding and Innovation Officer, to provide elite styling to match the in-person styling consultation. What did the app cost? By studying the CapEx contributions before, during and after the Ask Ralph launch to see what the cost of the app development was, there was a significant bump of Capex expenditure from $216 million in FY25 to $408 million in FY26 during the app’s development but the 10-K shows only $4.4 million of that increase is attributed to capitalized software. Nearly all of the increased spending sits in investments in physical retail: store buildout costs, and purchasing its SoHo and Newbury Street properties. Louvet stated in the earnings calls the brand remains committed financially to advance technology, AI and analytics.
The tool undeniably launched during a strong stretch for Ralph Lauren’s digital ecosystem, which saw sales accelerate from double-digit growth in Q2 to mid-teens rates in Q3 and Q4. Average unit retail (AUR) and comparable sales similarly surged. Yet management has consistently credited this momentum to strict pricing discipline, reduced discounting, and core marketing, leaving Ask Ralph’s true financial impact entirely uncredited.
What Microsoft is Doing for Guess
Ingesting Visuals: The system takes up to 12 product photos per SKU directly from Guess's PIM.
Extracting Implicit Attributes: Microsoft’s vision AI reads those photos like a human stylist would. For a single dress, it identifies details that aren't written down in the basic spec sheet like linen material, strapless neckline, pastel floral pattern, tropical vibe, and light fabric.
Generating Intent Vectors (The "Context"): Copilot Studio takes those visual traits and translates them into real-world consumer query scenarios. It tags the SKU with phrases like "outfit for a summer beach wedding in Greece" or "resort wear for hot climates."
Publishing Back to the PIM: It writes this machine-readable context layer directly into Guess’s backend database.
McMahon and I discussed editorial but from different meanings. I meant on the websites. Sue meant from elsewhere. Either way, Microsoft's multimodal catalog enrichment is using vision AI and LLMs to scan Guess’s product photos and generate high-density intent metadata such as 'wedding in Greece'. This allows AI shopping agents discover the item during natural-language searches.
Why the Industry Should Care
The Multi-Cloud Reality Check: What started as an all or nothing move to a single cloud provider is now multi-cloud. Enterprise retail tech stacks are a messy mix of Snowflake, AWS, Databricks, and Google is a pragmatic shift. It means tech strategy is finally adapting to how businesses actually operate.
The "Black Box" Backlash: Retail buyers and planners hated autonomous tools that make decisions behind closed doors. Validating that "autonomy" was a dirty word in merchandising, and shifting the focus to governed tools that show the work, touches on a major culture gap between tech vendors and retail operators.
Solving the LLM Discovery Problem: Most e-commerce sites were built for human eyes, full of brand storytelling, lifestyle photography, and fluffy copy. But AI agents shopping on a customer’s behalf can't parse that. Using generative AI to re-index product data into machine-readable facts is a critical fix for an entirely new channel of sales.
Zero-Party Data Loss: If a customer asks ChatGPT "What should I wear to a summer wedding in Greece?", the brand never sees that query. Bringing that conversational search in-house (like Ask Ralph) lets brands capture exact customer intent and prompt language for their own merchandising teams.
What Stood Out to Me: Don't Wait for Perfect Conditions to Start
The Tech Concept: McMahon advises retailers not to wait until their data is perfectly cleaned, unified, and organized across the entire enterprise before adopting AI. Instead, start with the specific decision you need to make right now and build support around that.
The Human Takeaway: Action Beats Preparation. People often paralyze themselves waiting for the "perfect" time to make a career move, start a project, or make a life change waiting until their resume is flawless or their schedule is clear. The lesson here is to launch with the assets you currently have, solve the immediate problem in front of you, and build the foundation as you move.
Coming Next:
Interview with JP MORGAN CHASE