SUMM

The concept of a "store" has evolved from clerk-assisted shopping to self-service big box retail, and then online e-commerce with vast selections
Early stores focused on personal interaction; the internet brought 24/7 global browsing but created a "sea of sameness"
Large online retailers still attract massive numbers of shoppers: Walmart gets almost 500 million, and Home Depot 170 million online monthly

AI shifts the focus from digitizing merchandise to digitizing participants (merchants and consumers) and their interactions
Static websites may be replaced by merchant and consumer agents, with payment infrastructure moving toward higher-level intent infrastructure
Future stores will feature dynamic, generative interfaces and infrastructure that adapt in real time to user and agent needs

Two possible futures: AI agents browsing optimized websites or programmatic access via APIs to merchants
Example: Consumer agents could handle intent expression, dynamic catalog browsing, and complete checkout on behalf of users
Programmatic access through unified endpoints could allow rich, on-the-fly interaction and purchasing without traditional web navigation
Focus remains on high-quality conversion and seamless transactions

Main transactional challenge: Enabling software to perform purchases reliably
Existing solutions: Virtual cards through platforms like Stripe SDK or delegated authentication (e.g. Visa's work), enabling agents to use real user payment methods
Buyer intent in the future could be explicitly captured through conversational data, not inferred from keywords or site metrics

Users coming from AI channels tend to have higher conversion, spend more, and have higher lifetime value
Easier returns for high-value users may reduce the cost impact of occasional mistaken purchases

Product pages today are static, but agentic commerce will require real-time availability, dynamic pricing, and product bundling across merchants
Existing solutions for inventory and store access (product feeds, web scraping) are inadequate; a unified API ("plaid for product data") is proposed

Today, buyer and seller preferences are siloed and narrow, with little data sharing
The future envisions expansive, two-sided preferences: richer context from both users and businesses, including real-time goals and inventory status
Challenges include the complexity and changing nature of preferences, conflicting buyer/seller interests, and incentive misalignment
Solutions may draw from financial market systems, introducing third-party market makers for coordination

True "agentic" commerce needs not only explicit intents and preferences but also intelligence that can reason and negotiate on behalf of both consumers and merchants
Infrastructure must evolve to coordinate and reason over participant needs and dynamically generate interfaces

Contrary to perception, Fortune 500 companies are often forward-looking due to surviving repeated technological shifts
Samsung example: From a fish merchant to a tech giant over 150 years, now planning for the AI-commerical future
Implementation steps: Abstracting product systems into cohesive APIs, connecting product data to other business data sources, integrating branding, enabling AI-driven interface experimentation, and supporting agentic transactions
Merchants see higher conversion rates from AI-driven buyers

Physical and digital stores will evolve, not disappear—returning to their origins as locations for conversation and transaction
The dominant interface of the future is likely to be a conversation

The shift to agentic commerce is happening quickly, with products like ChatGPT gradually integrating shopping but not yet completing the user journey in-app
Future sites may become modular, with data and components used across platforms rather than requiring users to visit websites
Payment rails: Credit cards are a practical bridge, but stablecoins/crypto could be a native long-term solution for agentic economies
Super apps (China, Brazil) parallel: AI platforms aim to become central commerce hubs, but merchants seek control over their presence and customer connection
Revenue sharing in the agentic commerce ecosystem is expected, likely through affiliate or high-quality data attribution rather than traditional advertising

Machines of Buying and Selling Grace - Adam Behrens, New Generation