January 15, 2026
Today's Briefing: January 15th 2026
NewsAgents Take
This briefing is a useful corrective to the current agentic commerce narrative. The long-term direction is clear and compelling, but the present reality is far more constrained than many headlines suggest. Consumer intent is already shifting in agentic ways, but the desire and capability to complete transactions autonomously remains limited. Payments sit at the end of the chain, but they are not the bottleneck. The real work lies upstream in trust, authentication, product data, checkout orchestration, and liability. Progress is real, but the gap between hype and operational reality remains wide.
A grounded conversation on the reality of agentic today
The conversation between Nikita Skitev and Colin Luce, CEO of payment orchestrator Basis Theory, is valuable precisely because it avoids overstatement. Colin is explicit that he is bullish on agentic and tokenised commerce in the long run, but equally clear that the industry is not yet close to widespread autonomous transaction execution.
One of the most important data points is behavioural rather than technical. OpenAI now sees somewhere between 800 million and 1 billion weekly users interacting with ChatGPT, processing billions of prompts. Around 2 percent of those prompts are commerce-related, which strongly suggests that intent formation is already happening in agentic environments. What is largely missing is evidence that users want those transactions to be completed autonomously, in-app, and immediately.
This distinction matters. Intent has shifted. Execution has not.
Colin also frames payments correctly as binary. A transaction either completes or it does not. But that does not make agentic commerce a payments problem. The real constraints sit upstream, including trust and authentication frameworks, machine-readable product catalogues, browser automation, and full-stack checkout logic that accounts for tax, shipping, and post-purchase flows. Until those layers mature, agentic commerce remains constrained regardless of payment capability.
The reality of the current hype cycle is sobering. Many implementations look closer to embedded checkout than true autonomy. Protocol proliferation is accelerating. Visibility and narrative are often rewarded almost as much as working systems. At a global scale where transactions are measured in the trillions per year, the number of genuinely AI-initiated transactions remains in the hundreds.
Early progress is worth recognising, but transparency about the remaining work is essential.
https://www.youtube.com/watch?v=wcOklRi2oUc
Understanding ACP versus UCP
Checkout.com’s breakdown of ACP and UCP is required reading for anyone trying to make sense of the current protocol landscape. Rather than positioning one as categorically superior, the article clarifies what each protocol is attempting to solve and where responsibilities sit.
For payments professionals, this is useful because it reinforces a key theme from the podcast. Protocols are necessary, but they are not sufficient. Without upstream clarity on identity, intent, catalogue integrity, and execution logic, protocol choice alone does not unlock autonomy.
https://www.checkout.com/blog/openai-acp-google-ucp-difference
Liability lessons from autonomous vehicles
The Progressive Policy Institute’s work on autonomous vehicles offers an instructive parallel. Even in an industry where failures can be fatal, liability frameworks have lagged capability. Responsibility is often diffused across manufacturers, software providers, and operators, with limited incentive for platforms to accept accountability.
The relevance for agentic commerce is direct. As systems begin to act on behalf of users, questions around misinterpreted intent, erroneous execution, and rogue behaviour will become unavoidable. The industry is not naturally incentivised to assign liability to AI platforms, but without credible accountability frameworks, trust will remain fragile.
Agentic commerce participants would do well to study how slowly liability has evolved in autonomous vehicles and plan accordingly.
