The Agentic Commerce Race: Battle for the Transactional Web

The transactional web is under siege. For nearly three decades, e-commerce has been dominated by discovery platforms and walled-garden marketplaces—Amazon’s search box, Google’s Shopping tab, and social commerce integrations. But the rise of AI agents is triggering the largest structural shift since mobile commerce. What’s at stake is nothing less than the operating system for online transactions.

OpenAI’s move into Instant Checkout (September 2025) was the spark. By embedding purchasing into ChatGPT, with over 700 million weekly active users and a conversion rate of 11.4%—double the benchmark of direct traffic—OpenAI has turned its conversational interface into a shopping mall. This isn’t incremental; it’s a redefinition of how transactions originate, flow, and settle.

The next phase is the protocol wars, a contest over who owns the rails of agent-driven commerce.

OpenAI / ChatGPT: Aggressive Offense

OpenAI has chosen offense over incrementalism. The Agentic Commerce Protocol (ACP), co-developed with Stripe, is open-source by design and integrates in a single line of code. This strategy mirrors the early internet: lower the barrier for merchants to onboard, create network effects around ease of integration, and monetize via transaction fees.

OpenAI’s early traction is impressive. Partnerships with Etsy and over one million Shopify merchants open the door to a long tail of unique, artisan, and structured-data-rich products that Amazon traditionally struggled to surface. By tapping into this ecosystem, ChatGPT transforms from a productivity assistant into a commerce intermediary.

Winners in this model are:

Merchants with structured data that agents can easily parse and present.Artisan and unique product sellers, who gain visibility outside Amazon’s ranking system.Early adopters who ride the first-mover tailwind.Payment processors like Stripe and PayPal, who benefit from transaction flow without owning the interface.

OpenAI’s real play is not just retail revenue—it’s becoming the transaction layer of the web. With ACP open-sourced, developers and merchants are incentivized to integrate, while OpenAI cements itself as the coordinating layer. In the first 12 months, it owns the speed advantage.

Google: Infrastructure Moat

If OpenAI is playing offense, Google is playing to scale. Its Agent Payments Protocol (AP2)—backed by Google and PayPal—leans on an unmatched moat: 50 billion product listings, refreshed 2 billion times per hour, coupled with decades of data across Search, Maps, and Gmail.

The advantage is personalization. Where OpenAI offers a protocol, Google integrates across its suite. AI Mode in Search already bridges discovery with transaction, and with AP2 it pushes checkout directly into the flow. For billions of users, this means their Gmail receipts, Google Calendar reservations, and Maps history all feed into commerce personalization at a scale no rival can match.

Yet Google carries legacy baggage. Its dependence on traditional ad-based discovery is vulnerable in an agentic world where users no longer scroll through pages of blue links or Shopping ads. Poor data infrastructure for merchant feeds, reliance on batch updates, and dependence on shelf placement all slow execution. Google will win in years one to three, but only if it can translate its infrastructure dominance into agent-native protocols.

Amazon: Defensive Position

Amazon, the incumbent giant, finds itself in an uncharacteristic defensive crouch. Its Rufus AI assistant is built to protect share, not to innovate on checkout. Amazon has so far doubled down on its walled-garden approach: discovery inside Amazon, checkout inside Amazon, with Prime as the stickiest part of the funnel.

The weakness is glaring: no checkout innovation. By 2025, Amazon’s profit projection of $700 million from this line looks modest in a landscape where transaction flow is up for grabs. If discovery begins to migrate toward agents—whether ChatGPT or Google’s AI Mode—Amazon risks losing the very traffic that fuels its engine.

In short, Amazon owns the warehouse but not the interface. Its discovery-only strategy could hold, but without embedding into the protocol wars, it risks being reduced to the supplier behind the agents.

Perplexity: The Aggressive Challenger

Perplexity is the nimble challenger in this race. Positioning itself as an “answer engine”, it has leaned into aggressive partnerships. The Buy with Pro launch (November 2024), followed by Firmly.ai integration (April 2025) and a PayPal partnership (May 2025), positions Perplexity as a serious transactional player. Its 5x growth in shopping queries proves there’s appetite for a multi-partner, open strategy.

Perplexity’s advantage is execution speed. With no legacy business to defend, it can move faster than both Google and Amazon. Its weakness is scale. Without a user base approaching OpenAI’s or an infrastructure moat like Google’s, Perplexity must ride alliances to stay relevant. Still, as an aggressive challenger, it forces incumbents to adapt and ensures the market doesn’t collapse into a duopoly.

The Protocol Wars: ACP vs AP2

At the core of this battle is the protocol layer. The question is simple: who defines the standard for agentic payments?

ACP (Agentic Commerce Protocol): OpenAI + Stripe, open-source, minimal integration friction. It thrives on adoption speed.AP2 (Agent Payments Protocol): Google + PayPal, ecosystem leverage, merchant scale. It thrives on infrastructure depth.

The competitive timeline is already visible.

0–12 months: OpenAI dominates as first mover.1–3 years: Google’s infrastructure and ecosystem begin to weigh in.3+ years: If standards fragment, the market risks becoming balkanized, with no single protocol layer achieving dominance.

History shows protocols matter. Just as TCP/IP and HTTP shaped the internet, the agentic commerce protocol will shape the transactional web. The player that secures merchant adoption at scale becomes the default transaction rail for AI agents.

The Bottom Line

The shift underway is the biggest e-commerce opportunity since mobile, but the window for adaptation is narrow. With 11.4% conversion rates, 45% YoY growth, and 700M+ weekly users, agentic commerce is not a speculative trend—it’s a structural reordering of digital trade.

OpenAI has first-mover speed and an open protocol play.Google has unmatched data and infrastructure.Amazon is defending its garden but risks losing discovery flow.Perplexity is punching above its weight with nimble partnerships.

The transactional web is no longer a marketplace of pages and search results. It’s becoming a protocol war, fought at the layer where agents decide what you buy, when you buy, and how you pay. Whoever wins the protocol battle doesn’t just capture e-commerce—they write the rules for the next era of the internet.

The post The Agentic Commerce Race: Battle for the Transactional Web appeared first on FourWeekMBA.