The following memo is a concept I have thought about for a while now having been working with brands and talking on their e-commerce challenges. User chat behaviour in a mobile format is not going away, now with ChatGPT, Claude etc we love more than ever to chat to our AI for advise, so why not for our shopping? This memo is open for an engineer willing to hack an MVP with me and up for grant allocation utilising the stack mentioned.
Nov 11 2024
The future of commerce isn’t about building more e-commerce websites with endless scrolling and overwhelming choices. On average 70% of online shoppers abandon their carts before completing a purchase. Brands are struggling with high cart abandonment, returns, and low sale conversion, largely because the traditional e-commerce model no longer serves them.
E-commerce has become a saturated space with uninspiring interfaces that fail to captivate. Current shopping behaviour includes browsing through Google search, Instagram, marketplaces and of course physical retail stores, second hand or thrift stores.
The problem extends beyond scrolling. The current online shopping flow is chaotic: when people turn to Google to find an item, they’re faced with thousands of search results across different websites, leading to fatigue and frustration. Many shoppers spend hours navigating product pages and adjusting search parameters just to find something that fits their needs.
In a recent survey by Accenture (June 2024) involving 19,000 consumers worldwide, 74% reported abandoning an online shopping cart at least once in the past three months. They felt “bombarded by content, overwhelmed by choice, and frustrated by the effort required to make decisions.” This indecision was especially common in categories like clothing (79%), cosmetics (79%) and flights (72%).
Consumer shopping habits are shifting toward personalisation and exclusivity. People are tired of algorithms and recommendation engines that make them feel like everyone else. The shopping experience is ripe for an upgrade: rather than browsing endless product pages, shoppers are now engaging with brands in more social and interactive ways.
Today, people are accustomed to chatting on mobile apps and getting instant insights from AI tools like ChatGPT and Claude. Imagine this behaviour applied to shopping—a commerce app with a personal AI agent that understands your taste, style, and preferences, offering curated recommendations for clothing, furniture, beauty products or brands that align with you. This would make shopping more enjoyable and personable, eliminating the need to scroll endlessly through websites or social media feeds.
Source Accenture report 2024
Ok but what are some of the limitations and surely Chat can already do this?
One of the obvious hurdles for a personal shopping agent lies in autonomous payments. Current financial infra assumes human agency in transactions for obvious reasons. ‘Paywalls, two-factor authentication, CVV verifications, and dynamic security checks are designed to confirm human intent, creating barriers for AI agents.’ - Claude. Basically blocking any bot or AI driven transactions.
Marketplaces and payment gateways are not yet architecturally prepared for machine-to-machine commerce, where an AI agent could seamlessly browse, select, and purchase on a user's behalf without constant human intervention. All that said, clearly a barrier to entry however new developments could help solve this, I’ll get to that later!
Whilst yes Chat can give you recommendations and deals on products via Shopping GPT, the UX is still limited and not autonomous because of restrictions outlined. Existing examples include ‘Shop GPT’ available on Shopify and dodgy looking apps likes ShopGPT.
Shopping GPT has some sass!!
Why I think this creates such a lucrative business model and a vast design space opportunity?
AI agents can serve millions of users simultaneously, eliminating the need for extensive human coordination and labour in brand distribution and attention capture.
Several promising business models emerge: tiered subscription-based personal shopping assistants, brand partnership commissions, and premium ad placement for sponsored products. The right model depends on the specific product and target audience. For enhanced personalisation, brands, designers, and stylists could offer their expertise as premium features for a more personalised experience. All that data can then be fed back into the Agent for refined customer preference data.