Building an AI-Native Marketing Team, AI in Online Shopping, and Burned-Out Tech Workers

Set the example.

Good morning

In today's edition, among other things:

• AI is revolutionizing online shopping

• The AI playbook that saved ElevenLabs $100k+

• Building an AI-Native Marketing Team

• Tech Workers: Burned-Out but Building

• The Great ERP Unbundling

• Is the EU AI Act Actually Useful?

Onwards!

---

AI is revolutionizing online shopping

There’s not a week that goes by when I don’t get a pitch deck with an AI for your wardrobe startup. On one hand, the bar is building lower than ever; on the other, it signals a significant transformation in online shopping.

A decade ago (and still for some companies), “shopping online” meant typing keywords, scrolling endless thumbnails, and hoping the size chart wasn’t lying. Today the keyword feels outdated. AI is turning commerce from a treasure-hunt like experience into a concierge experience that anticipates, personalizes, and negotiates on your behalf (if done right).

Here’s a16z:

Online shopping is becoming intelligent, predictive, and visually intuitive. Soon, the right products, at the right price and size, will find you, not the other way around.

OpenAI is starting to treat shopping as a dedicated use case, adding tools like Operator, which lets users ask for product recommendations, and Deep Research, which aggregates specs, reviews, and price data for side-by-side comparisons. Meanwhile, startups are building vertically integrated solutions for every step in the stack, from product discovery and smarter, multimodal search to virtual try-ons and autonomous post-purchase support.

If you “walk into” any e-commerce feed and you face the “paradox of choice”: infinite SKUs, questionable reviews (I don’t trust them per se), and prices that change hourly. It’s different with AI driven commerce:

• OpenAI’s new Operator tool and counterparts like Deep Research parse specs, reviews, and price histories in one query, then deliver a ranked shortlist instead of 4 000 results.

• Start-ups are slicing the funnel vertically—some own multimodal search, others own post-purchase support—so the consumer moves from overwhelm to one-shot ask.

• The interface is no longer a catalog; it’s more of a conversation (do not confuse with chatbots). The best answer might be a single product, a generated comparison table, or simply “wait a week—prices are falling.”

AI Mirrors

Trying clothes on a rectangular screen used to demand imagination (and free returns). Diffusion models erase that leap.

• Doji and similar platforms build a LoRA avatar from a couple of selfies; seconds later you see the drape of a blazer on your shoulders, not a size-zero mannequin.

• As 3-D generation improves, fit and movement—how denim creases at the knee, how silk hangs when you walk—will preview in real time.

• When the mirror speaks truth, return rates and wardrobe anxiety plummet. Retailers save on logistics; shoppers save on guilt.

AI Stylists

Owning clothes is easy; combining them is art. Half of what sits in U.S. closets goes unworn.

• Digital closets like Alta ingest what you already own, today’s weather, and tomorrow’s calendar to suggest outfits—plus the missing belt that would pull it together.

• Over time, the AI learns your “style vectors”: maybe you brave color at brunch but stick to navy for investor day.

• Outfits, once a messy, context-dependent idea, become first-class data objects the system can remix infinitely.

AI Inventory

Customization once meant picking a font and button colors. Generative pipelines now let the consumer co-design from scratch.

• On Arcade AI, typing “star-shaped ring for a birthday party” doesn’t yield listings; it spawns original CAD-level designs with live pricing.

• Because manufacturing is shifting to modular, on-demand, and even 3-D-printed flows, the SKU can be produced only when ordered. No inventory risk; no overstock landfill.

• Retailers harness forecasting models to predict which co-creations will trend and pre-stage just the raw materials, shrinking working capital.

Discovery and Search

Deal-seeking is a universal customer behaviour; AI just does it faster.

• Tools like Dupe reverse-image that $2 600 designer sofa and surface visually similar options under $2 000—or under $200—without sacrificing style.

• Beni pivots the same pipeline toward sustainability, surfacing second-hand alternatives for every fast-fashion impulse.

• Price becomes a moving target the bot tracks continually, nudging you when the curve dips in your favor.

Customer Service

Support has long been the cost center of retail. LLM-powered agents such as Decagon flip it into a differentiator.

• These bots update shipping info, process returns, and answer the “Where’s my refund?” email instantly, driving higher deflection rates and faster first replies for brands like Curology.

  • Zowie.ai from Poland is one of many examples.

• Freed from rote tickets, human reps handle edge cases and relationship-building; the AI escalates only what truly needs empathy.

What’s ahead: personalized, predictive, AI-driven shopping

AI-driven shopping is still in its early stages, but the opportunity is vast. Startups are approaching the opportunity from different angles, improving search and discovery, enabling virtual try-ons, offering customized inventory, and organizing wardrobes to inspire new outfits and purchases. These early innovations are laying the groundwork for a fully integrated personal shopper that understands your style, remembers your preferences, and grows smarter with every interaction.

Imagine an assistant that restocks your daily essentials, anticipates your needs, recommends your next purchase, and helps manage all your purchases in one place. With each interaction, the assistant learns through a personal data flywheel, drawing on your browsing habits, chat history, and purchase patterns to deliver a hyper-personalized experience. As it accumulates more context, it removes friction from the shopping process, eliminating the need to toggle between shopping websites or maintain manual lists.

The shopping mall was a place; the e-commerce site was a page; the AI shopper is a 24/7/365 “presence”. It compresses discovery, evaluation, customization, and service into a single dialogue—one that should get smarter each time you interact.

---

The AI playbook that saved ElevenLabs $100k+ and helps them ship daily