This paradox was suspiciously absent in most of my conversations at Money20/20 last week. Here's what many investors are missing: everyone's building consumer shopping agents. Relatively few are building for merchants. That asymmetry creates opportunity.
Agentic commerce carries enough hype to make any skeptical investor wary, but early data suggests this time is different. Adobe reports that traffic to US retail sites from GenAI browser and chat services has surged by an astonishing 4700% as of July 2025. Crucially, these are not low intent shoppers! They spend 32% more time on-site, browse 10% more pages, and have a 27% lower bounce rate compared to typical visitors (Adobe). Consumers are ready for AI; merchants are not. That is a problem worth solving.
It’s critical to recognize that AI agents interact with the digital world in fundamentally novel ways. For them to transact reliably and securely on our behalf, a new infrastructure stack must emerge to enable a rich, autonomous ecosystem unfettered by human lag and limitation.
In this light, three investment areas stand out to me:
1. AI agents to help facilitate considered purchases
There is a clear opportunity to build AI agents designed for high-consideration shopping. AI excels at the most tedious parts of shopping: comparing options and synthesizing information. That makes AI perfect for considered purchases where confidence and comprehension matter more than speed: insurance, enterprise software, home improvement, mattresses, luxury goods, healthcare services.
Consumers today rely on scattered reviews, outdated blogs, and forum threads. Agents that can extract insights from unstructured sources (e.g. video reviews, PDFs, manuals, Reddit) and compress days of research into a coherent recommendation will redefine product discovery.
Amazon mastered endless selection. DoorDash mastered speed. The next retail giants can be built by mastering trust and expertise through AI-driven consultation. Startups positioned here will look more like vertical SaaS companies than marketplaces, blending data aggregation, domain-specific LLM fine-tuning, and agentic UI design.
2. Merchant enablement and protocol infrastructure
The most overlooked opportunity lies in merchant enablement. No matter which consumer shopping agent dominates, merchants will need infrastructure to accept, interpret, and verify agent-driven transactions.
A protocol boom is already underway. At Money20/20, PayPal introduced Agent Ready and Store Sync. Stripe and OpenAI co-developed the Agentic Commerce Protocol (ACP). Google announced the Agent Payments Protocol (AP2). Visa and Mastercard followed with their own Trusted Agent Protocol and agentic commerce frameworks. Even startups in payment infrastructure, such as VGS (a Vertex portfolio company), have articulated what “agent-ready” merchant architectures will require, emphasizing secure data exchange, tokenization, and identity abstraction as foundational layers for agentic transactions.
This standards explosion is confusing for merchants and, thereby, creates fertile ground for startups. We’ll see demand for protocol-alignment tooling, merchant SDKs, certification platforms, and trust registries.
In parallel, merchants will need their own branded agents to help them negotiate with buyer agents, personalize pricing, and retain control over customer relationships. Just as Shopify enabled small merchants to transact online, a new generation of startups will help them transact with AI.
3. Trust, identity, and authentication layers
Credit cards unlocked catalog shopping. SSL enabled e-commerce. Fraud scoring made marketplaces viable. Each commerce revolution needed its own trust layer. Agentic commerce is no different.
The challenge here is familiar: when an agent places an order, merchants face the classic double verification challenge. Is this agent real (i.e. authentic)? Is it authorized to spend on behalf of this specific user? Yes, this sounds like old news, but our authentication and authorization stacks today are built for humans! Visa's Trusted Agent Protocol and Mastercard's Agent Pay are early responses to address part of this, but only part.
The real, near-term opportunity sits in the gaps: agent identity registries, zero-trust verification systems, reputation systems, behavioral signatures, and escrow-style authorization frameworks. As with payments, whoever controls authentication will control data, distribution, and ultimately, economic leverage as the trust rails for agent-to-merchant interaction.
For founders building in agentic commerce, the setup is exceptional: explosive consumer traffic growth, a merchant readiness gap, and critical trust infrastructure still up for grabs. Agentic commerce has only just started to change the way we think about shopping, and the builders who enable this ecosystem will define the next era of commerce.
If you're building here, I want to talk!
Originally published on LinkedIn.