Something fundamental is changing about how people buy things online. The shopping journey that once required consumers to visit multiple websites, read reviews, and manually compare options is being handed over, stage by stage, to autonomous AI systems. These agents can research products, apply filters, and assemble purchase-ready selections without any active participation from the buyer. And as this transition accelerates, the part of the retail process that happens last, checkout, is becoming the part that matters most.
Rokt has built its business around this understanding. The company’s analysis of what the industry calls agentic commerce reflects a clear view: that as AI systems take over the discovery and evaluation phases of shopping, the transaction moment becomes the primary site of human engagement, commercial opportunity, and strategic differentiation.
The macroeconomic context gives weight to that view. McKinsey research published in October 2025 estimated that agentic commerce could drive up to $1 trillion in orchestrated U.S. retail revenue by 2030. Global projections from the same research reach as high as $5 trillion. The consulting firm placed this shift in the same category as the web and mobile revolutions, while noting that the adoption curve could be steeper.
Major technology companies have already moved quickly to establish positions in this space. OpenAI’s Operator tool allows users to automate purchasing tasks through ChatGPT. The company also partnered with Stripe to launch an Agentic Commerce Protocol that enables in-chat transactions. Google, Amazon, PayPal, and Mastercard have each introduced their own versions of autonomous shopping tools.
What all of these systems have in common is that they bypass the touchpoints retailers have traditionally used to influence consumers’ search results, category pages, product detail pages, and review sections. When an AI agent handles those stages, the first opportunity for a retailer to engage a consumer directly is at checkout. This is what makes checkout architecturally significant in ways it has never been before.
Rokt’s analysis describes checkout as the validation layer between algorithmic recommendation and consumer approval. Even consumers who delegate most of their shopping to AI agents want to review and confirm purchases before completing them. Research cited in Rokt’s work shows that people want the ability to pause, adjust, and reverse agent-driven decisions. Checkout is where that authority is exercised.
For this to work effectively, checkout must be adaptable. A shopper arriving after a highly AI-managed purchase journey has different expectations and needs than one who has browsed manually. Context-awareness responding to signals about how the purchase journey unfolded, what level of agent involvement occurred, and what the individual’s preferences suggest is the difference between checkout that reinforces trust and checkout that undermines it.
Rokt’s technology infrastructure is purpose-built for this requirement. The company’s Rokt Brain processes more than 1.95 trillion data points annually across 7.5 billion transactions to determine the optimal experience for each individual at the point of purchase.
The results of this approach translate into measurable commercial outcomes. Rokt Pay+ produces $400,000 in incremental profit per million transactions. Rokt Thanks delivers up to $500,000 per million transactions on confirmation pages. Rokt Ads achieves a 4.03% global click-through rate, ten times the performance of Google Display advertising benchmarks.
Retailers who treat checkout as a strategic priority and invest in flexible architectures and real-time personalization will be equipped to capture value as agentic commerce scales. Those who continue to treat checkout as an afterthought risk becoming irrelevant to the AI agents that increasingly determine where and how consumers purchase. Rokt’s role in this transition is to provide the infrastructure that makes checkout worthy of the moment it now represents.
