Tech • AI • Robotics • Game
People are making €100M+ in sales on this app
5/10TL;DR
E-commerce is splitting into two fast-growing models: highly emotional live shopping led by Whatnot, and highly rational purchasing handled by AI shopping agents, setting up a battle over who controls access to the end customer.
KEY POINTS
Two opposite buying behaviors are emerging
Online retail is moving in two directions at once. On one side, shopping is becoming entertainment, driven by impulse, community and urgency. On the other, repetitive purchases are likely to be delegated to AI agents that compare price, quality and availability with far less emotion than human buyers.
Whatnot turns shopping into a live spectacle
Whatnot combines the mechanics of TikTok Live and eBay auctions. Users scroll through live video streams where sellers present products, answer comments, trigger countdowns and launch instant auctions. The design compresses decision time and encourages immediate purchases, often after buyers have already entered payment details and received a signup credit of around €10 to €15.
The platform is built for impulse spending
The selling sequence is highly engineered: a low starting price, a short timer, visible bidding activity and a charismatic host create fear of missing out. In some cases, products are opened live after purchase, especially in categories such as Pokémon cards, adding suspense and gamification. The result is a retail format closer to a show than a catalog.
French sellers are already posting large numbers
The model is no longer marginal in France. A French trading-card seller reportedly generated nearly €700,000 in sales in 18 months. Another seller moved €60,000 of Pokémon cards during a single live session. A luxury-bag seller reached €50,000 in sales in three hours, while some merchants now say 30% to 40% of monthly revenue comes from the platform.
Influencers are becoming high-pressure retailers
The most effective hosts are not traditional sales clerks but creators who mix entertainment, intimacy and persuasion. They joke with audiences, answer questions, build familiarity and then sell. That makes live commerce especially powerful for fashion, collectibles, luxury accessories and other products where desire, rarity and social proof matter more than pure utility.
A new service economy is forming around live commerce
The rise of Whatnot creates openings for agencies, liquidation specialists and outsourced sales teams. Brands with unsold stock, returns or niche inventory can use trained live sellers to move goods quickly. Businesses can also emerge to manage live operations for labels that lack on-camera talent, audience-building skills or auction expertise.
AI shopping agents could reshape the other half of retail
For routine purchases such as groceries, household goods or school supplies, shopping agents promise the opposite experience. Instead of browsing, users will describe a need and let software choose the best option based on price, quality, delivery and preferences. If that behavior scales, traditional product discovery on marketplaces could weaken sharply.
Control of the entry point is becoming strategic
The biggest prize is not the sale itself but control of the interface through which purchases begin. If consumers rely on Meta, Amazon, OpenAI or other assistants to choose products, those companies gain leverage over both merchants and customer relationships. That could reduce the value of current advertising slots and replace them with new forms of paid placement inside agent recommendations.
Three business models stand out for AI intermediaries
Shopping agents could monetize retail in at least three ways. They can sell sponsored ranking inside recommendations, take a commission on conversions like a digital salesperson, and charge fees on transactions flowing through their systems. For platforms operating at scale, those models could become more important than subscription revenue alone.
Brands may need to market to algorithms, not people
If agents handle more purchases, merchants will have to optimize product data with machine-readable clarity. Price, quality claims, availability and differentiators will need to be explicit and structured so AI systems can compare them accurately. In that world, weak descriptions or incomplete data could mean invisibility, even for otherwise strong brands.
CONCLUSION
Retail is increasingly dividing between purchases made for excitement and purchases made for efficiency. The companies that master either emotional engagement or algorithmic access, and especially those that control the customer gateway, stand to capture the next major gains in e-commerce.
Ask a question