Scaling BPO success with agentic AI: Real-world client wins -
Watch now
The conversations that stood out most and what they mean for your Holiday 2025
The energy at eTaili Boston’s AI Summit was special this year. Retailers weren’t debating whether or not AI belongs in their space.
Everyone was trading notes on what’s actually working.
We spent three days listening to teams from Wayfair, Lowe's, CVS, and dozens of others. One thing was clear: AI is shaping marketing, merchandising, operations, and commerce in very real, very tactical ways for retailers.
Here’s what stood out most, and what it means for Holiday 2025 in just a few months.
SEO isn’t enough anymore. Your customers are asking ChatGPT, Perplexity, and Gemini to find products, not just Google.
Answer Engine Optimization (AEO) makes sure your products show up when AI engines answer questions like, “What’s the best skincare subscription box under $50?”
What it means: Start feeding product and brand data into these engines now. The more structured the better. Think of this like SEO in 2010, when early movers owned the channel.
The biggest AI leverage in retail isn’t “sexy.” It’s functional. Operations like inventory management, promotions, bundles, and refunds.
One retailer shared how they use AI to avoid overselling across TikTok, Etsy, and Shopify, while another explained how they AI-automated their entire refund approval process.
IPSY knows it well. Their team used to work nights and weekends making pricing changes. Now they handle 800+ products in minutes.
What it means: Start with the manual work that makes your team want to quit. Automate that first.
Retailers are cutting weeks off creative cycles with AI product imagery, quick-turn videos, and mockups that used to take designers days.
But here's the catch: Writing 1 PDP with AI is easy. Writing 10,000 is hard. Scale breaks most AI tools because of context limits and integration needs.
What it means: You need tools that actually connect to your commerce stack, not standalone content generators.
While 80% of first-touch customer service can be automated, high-emotion or high-value interactions should stay with human operators.
It might look something like this: AI handles order lookups, FAQs, and routing. Then human operators step in for brand-sensitive issues, and AI assists with suggestions for consistency.
What it means: Map your customer service workflows and identify which interactions truly need human touch versus those that just need speed.
This holiday season will mark the first major wave of consumers using AI engines for discovery, especially for gifting and personalized recommendations.
While full checkout inside AI engines may not dominate yet, discovery will. Which means retailers’ product catalogs need to be structured, tagged, and AI-crawlable now.
What it means: Audit your product data structure. If an AI engine can't understand your catalog, your customers won't find you.
Messy data kills AI ROI. Before you chase a new tool, retailers need to clean up product SKUs, descriptions, and tags.
Fortunately, you can use AI itself to accelerate your cleanup. Multiple retailers shared how they used AI to standardize thousands of product descriptions in just days.
What it means: Start with data quality, then everything else builds on this foundation.
No "perfect" AI tool exists yet. Retailers should continuously test different tools across categories: merchandising, content, commerce, and operations.
Tools vary widely in impact depending on vertical. What works for fashion might fall flat in home goods.
What it means: Adopt a portfolio approach to test multiple tools with small budgets and clear success metrics.
Several retailers shared horror stories about AI-generated content that infringed on logos or created damaging brand imagery.
To avoid this, build internal policies around acceptable AI use: no logo replication, no celebrity deepfakes, and clear review processes for generated content.
What it means: Write your AI content guidelines now, before your team needs them in a crisis.
The old funnel is dead. TikTok Shop and AI recommendations mean customers can jump from awareness to purchase in seconds.
Nicole, VP of Program Management at IPSY, lived it firsthand: "Non-technical people can now leverage APIs and browser automation versus waiting on engineering."
What it means: Your teams need to move as fast as your customers do.
Real enterprise AI means integrating your commerce, ERP, and marketing stacks, making decisions on when AI should be creative versus predictable, and aligning IT, legal, and security teams before rollout.
IPSY's CTO Sree Sreedhararaj nailed it when he said, "Don't think, 'How do I invest millions and change the world?' It's not going to happen overnight."
What it means: Start with one workflow that hurts, prove value, then scale systematically.
You don’t need the flashiest AI to move the fastest. You’re better off cleaning up your data, offloading your operations, and preparing to be discovered in AI engines.
Holiday 2025 is three months away. Are you ready?
See how enterprise retailers are building AI agents that handle complexity while you focus on strategy and growth, and book a demo to start building your first one.
Expanding possibilities for people at work
