In this episode of The Orita Podcast, Aaron Schwartz sits down with Jennifer Peters, Director of D2C Marketing Technology & Digital Compliance at Olly, and Avani Oswal, Head of Digital at Elizée & Founder of Catalyst, to unpack what it actually looks like to build and scale a modern commerce brand in the age of AI, whether you're a category-leading CPG or an emerging luxury label finding your footing.

Jennifer brings nearly a decade of Shopify development experience and a front-row seat to omnichannel complexity inside a Unilever-owned brand. Avani has lived multiple sides of the ecosystem, from Macy's to Ipsy to GameTime, and now leads digital for a high-end Italian shoe brand while running one of commerce's most tight-knit operator communities. Together, they offer a candid, operator-level view on what's actually changing, what isn't, and where to place your bets right now.

They dive into:

• Why the last six weeks have changed more than the last few years and what that means for discoverability

• GEO, AEO, and LLM visibility: separating real strategy from AI theater

• Why great content still wins and the danger of creating content just for bots

• How to pilot new tools without overcommitting: the case for breaking things before you buy

• The tools they're actually using for product data enrichment, AEO measurement, and AI-generated imagery

• What to automate (supply chain, campaign optimization, photo shoot extensions) vs. what still needs a human

• Personas in flux: why hard-and-fast customer profiles may be doing more harm than good

• Omnichannel loyalty strategy: receipt scanning, direct mail, and why you should start before you're ready

• Why TikTok Shop customers need a totally different retention playbook and how direct mail is winning there

• "Here for the customer, not for the channel" building internal structures that don't pit teams against each other

• Luxury loyalty: what incentives actually work when your customer earns enough that 10% off means nothing

From AI strategy and tool evaluation to loyalty, attribution, and omnichannel operations, this episode is packed with practical insight for operators trying to build brands that last, whatever channel their customers choose.

👥 Meet the guests

Jennifer Peters, Director of D2C Marketing Technology & Digital Compliance at Olly

Avani Oswal, Head of Digital at Elizée & Founder of Catalyst

🎙️ Hosted by Aaron Schwartz, Co-Founder & Co-CEO at Orita.ai

🎧 Chapters

[00:00] Meet the Operators: Shopify Dev Turned Brand Leader & Luxury Shoe Disruptor

[02:00] From Italian Heels to Supplement Aisles: Building Brands That Last

[05:24] Why Agency Life Breaks You, and Makes You Better

[09:09] "The Last 6 Weeks Changed Everything": AI's Real Impact on Commerce

[11:12] How Smaller Brands Are Beating Incumbents With AI Right Now

[16:30] Trust But Verify: Using AI Tools Without Getting Burned

[20:46] Stop Doing AI Just to Tell Your Board You're Doing AI

[23:05] The Right Way to Pilot a Tool, and Why "Break It" Is a Green Flag

[27:04] Is Your LLM Traffic Being Undercounted? The Attribution Problem Nobody's Solving

[29:32] How to Run an Experiment That Actually Means Something

[32:36] What to Automate, What to Protect: The Human vs. AI Breakdown

[34:24] Your Personas Are Lying to You: Here's What to Do Instead

[38:15] Why Warehouse Tech Is Living in the Past, and Who's Going to Fix It

[40:14] "Here for the Customer, Not the Channel": The Omnichannel Mindset Shift

[44:07] Direct Mail, TikTok Shop, and the Loyalty Tactics Actually Working Right Now

🌐 Check out Orita.ai for more episodes of The Orita Podcast, where we sit down with the most interesting operators, marketers, and founders in e-commerce to talk about what’s working right now. Hosted by the Orita team. Orita uses machine learning to help brands send smarter, more profitable marketing, by figuring out who actually wants to hear from you and when.

🚀 Get your free email audit from Orita, turn segments into superpowers: https://orita.ai

🎧 Listen on Spotify & Apple

📩 Subscribe for more founder‑grade retention intel