DMs Are the New Storefront: AI Messaging That Gets Gen Z to Buy

Image from: freepik.com

AI Style Assistants in Fashion Retail: From Product Discovery to Purchase Nudges

“Gen Z doesn’t read emails. They expect brands to talk to them like a best friend - in DMs, not newsletters.”

This is the new reality for fashion brands. It’s not enough to have great products - you need to speak the language of your audience, where they are, and how they like it.

A leading international fashion and home décor retailer, with presence in over 30 countries, knew it was time to reimagine how they connect with their customers. While they had millions of users engaging via app, email, and website, their messaging capabilities - especially on chat-native platforms like WhatsApp - were lagging behind.

They started by using WhatsApp to send discount codes. But they had bigger ambitions.

The Challenge: Move from Static Notifications to Smart Conversations

The retailer wanted to:

• Proactively engage users with birthday messages, style advice, cart nudges, and restock alerts
  
  

• Extend messaging across existing channels - app, email, web chat
  
  

• Build behavioral profiles to personalize at scale
  
  

• Automate loyalty and referral journeys
  
  

• Reduce customer service load with 24/7 conversational support

And all of this had to speak fluently in a Gen Z tone.

Our Approach: A Modular AI Messaging System for the Gen Z Era

We implemented a full-stack solution - powered by LLMs, Retrieval-Augmented Generation (RAG), and Visual Language Models (VLMs) - tailored to the client’s global catalog and young customer base.

1. User Intelligence Engine

Unified data from CRM, web, and app to create rich behavioral profiles - capturing preferences, engagement patterns, and style inclinations. These powered personalized interactions across the board.

2. Campaign Orchestrator

Automated decision logic chose the right message type (e.g. abandoned cart, birthday, drop alert) and the best time and format for delivery - personalized to each user’s journey.

3. Message Composer for Gen Z

Using fine-tuned LLMs, we developed a tone-appropriate copy engine that could generate punchy, emoji-friendly, on-brand messages that resonated with Gen Z without sounding forced or robotic.

4. Style Assistant: Smart Conversations Start with Smart Metadata

One of the most ambitious components was the AI Style Assistant - a chat-based product discovery tool that helped users ask questions like:

• "Will this top be see-through?"
  
  

• "Does this dress have a tight fit?"
  
  

• "Can I wear this coat in snow?"

But there was a problem.

Poor Metadata = Poor Chat Experience

Most product catalogs weren’t built for natural language understanding. Critical attributes like fit, lining, texture, and seasonal suitability were either missing or inconsistent.

“The tone was right. The answers weren’t.”
That’s when we launched a full-scale metadata enrichment initiative.

Metadata Enrichment Project: Turning Product Data into Product Conversations

We enriched the catalog through:

• Visual Language Models: Automatically extracted features from product images (e.g. sleeve length, fabric thickness)
  
  

• Text Parsing: LLMs analyzed descriptions to standardize and normalize language
  
  

• Schema Expansion: Collaborated with the client to define new metadata fields (e.g. “lining presence,” “climate suitability”)
  
  

• Feedback Loop: Used user queries to identify gaps and continuously improve coverage

5. Loyalty & Referral Automation

We built plug-and-play workflows for refer-a-friend campaigns, milestone rewards, and reactivation nudges - driven by user behavior and peer engagement patterns.

6. Conversational Support Automation

24/7 support powered by a RAG model combined with the enriched catalog, internal help docs, and fallback routing. The assistant could now respond to both “Where is my order?” and “Can I machine wash this?” - all in the same session.

“Our support queries dropped 30% within weeks.
But more importantly, customers started saying ‘This feels like texting a friend.’”
– Head of Digital Experience

Use Case Highlights

• Style Assistant: Product advice in Gen Z tone with accurate visual + text-based data
  
  

• Back-in-Stock Alerts: Triggered by product interest and wishlist actions
  
  

• Abandoned Cart Nudges: Personalized messages with visual previews and incentives
  
  

• Birthday Campaigns: Fun, timely greetings that led to spikes in conversions
  
  

• Multi-product Offers: Bundled suggestions based on style and purchase history

The Results

• +50% projected increase in monthly active messaging users
  
  

• 1.9M+ personalized conversations per month
  
  

• 24/7 smart automation for support and product discovery
  
  

• Scalable architecture, already rolling out across 29 markets

From Pilot to Playbook

This localized implementation has become the brand’s AI messaging blueprint - flexible, modular, and future-proof.

Want to Build AI Messaging That Talks Human?

At Nuefunnel, we design modular AI systems that turn catalogs into conversations - and interactions into loyalty.

Let’s talk.
Whether you're looking to enrich your metadata, personalize messaging, or automate support, we’ll help you get there - chat by chat.