AI & Agentic Commerce

AI and Agentic Commerce - The Rise of the Machine Economy

Takeaways from Peebles' Q2 2026 research programme: 60 interviews with leaders across agentic AI, tokenisation, and security.

After leaving Consensys at the beginning of May, I decided to conduct research into the agentic AI, tokenization, and security spaces. It was originally done as a way to help me identify new areas to focus my own time on. After a few meetings I realized there was inherent value in the content and decided to expand the process - I ended up speaking to 60 leaders in the space. The research is divided into two main sections: my own synthesis of the 60 in-person interviews, and an AI-based synthesis of the notes using a multi-stage tagging and memo-writing process with Perplexity Comet's research capabilities. Here are some of my own takeaways. Download the full Q2 research report here.

The Rise of the Machine Economy

It is becoming apparent that AI requires the re-writing of the internet. Like 1998, when every company needed a website, or 2008, when every company needed an iOS app - every company now needs to set up their web presence to be agent discoverable and consumable.

On the competitive front, businesses need to focus on moats that AI cannot replicate or access. The four key moats required to withstand new AI-based businesses are:

• Distribution and access to users
• Liquidity or AUM that AI cannot get access to
• Integrations, contacts, and partnerships that AI cannot reproduce
• Proprietary data that AI does not have access to

"Anyone with a UX/UI moat will disappear quickly." Sarah Fryer, CFO of OpenAI

In the research, a CEO of a major web3 venture studio stated that crypto is a perfect system for agentic commerce, allowing for proof of ownership - and that we should expect machine banking, and more agents than humans.

An agentic API platform provider mentioned the need to enable both sides of agentic commerce:

1. Enabling API providers to be agent-native via agent authentication and agentic payments, so the website itself is consumable by agents; and
2. Enabling developers via observation, controls, and a unified way to manage and pay for API consumption.

One agentic commerce provider mentioned key use cases for agentic commerce via stablecoins, such as passports for agentic shopping, API consumption (for example, cheaper GPUs), agentic trading of crypto or stocks, and B2B procurement.

The Impact of AI on Security and Identity

The two other areas heavily impacted by AI are security and identity.

• Security - AI has made hacking infinitely easier and cheaper. Projects and organizations are having to respond with "hyper secure" strategies. The security industry is responding with security-at-scale solutions like vulnerability mining marketplaces, expert-AI-triage workflows, and AI-assisted-but-human-verified pipelines.
• Identity - AI is breaking the verification layer of the internet. With deep fakes and fake IDs, it will eventually impact the verification layer of society. We see the identity industry responding with verification-at-scale solutions such as zkTLS-based documentation verification and AI-based KYC processes - and we believe that zero-knowledge proofs can restore that verification layer.

For security, I've published a follow-on article here. You can also check out the full Q2 research here.

For identity - a topic of interest to me ever since I started using EVM wallets back in 2017 - this will be a focus of the research in Q3. If you're in the KYC, compliance, or identity space and would like to chat, please message me or email ian@peebles.co.