Big tech and payments firms are taking agentic commerce seriously

Executives from PayPal and Google Cloud said at Consensus Miami that autonomous AI-driven commerce will likely require open payment systems, programmable money infrastructure and machine-readable transaction layers to scale effectively. According to CoinDesk, the discussion focused heavily on how AI agents could eventually transact directly with services, merchants and each other using crypto-based rails.

The idea is that AI agents may become active economic participants online, purchasing services, accessing APIs and coordinating transactions autonomously. Traditional payment systems were largely built for humans, while programmable blockchain-based systems may fit machine-to-machine commerce more naturally.

Open payment protocols are seen as essential

One of the main themes from the discussion was interoperability. CoinDesk reported that speakers emphasized the need for open payment protocols rather than isolated closed ecosystems.

That matters because autonomous AI agents may need to interact across many different platforms, applications and service providers simultaneously. Open standards make that kind of coordination far easier than fragmented proprietary payment networks.

Crypto infrastructure already leans heavily toward open protocols and programmable interoperability, which is one reason the sector keeps appearing in these AI-commerce discussions.

Machine-readable merchant catalogs could become a core layer

Another interesting point from the panel was the importance of machine-readable merchant catalogs. In simple terms, AI agents need structured digital environments they can understand programmatically.

Humans can browse messy websites visually and interpret context intuitively. AI agents work differently. They need standardized, structured information formats to discover products, compare offers and execute purchases autonomously.

That means future commerce systems may need to evolve beyond traditional storefront design toward data structures optimized for software interaction.

Multi-party custody is another major challenge

The panel also highlighted custody and authorization as critical issues. CoinDesk reported that multi-party crypto custody systems may become important for safely managing AI-driven financial activity.

That is because autonomous AI systems handling payments create obvious security risks. If an AI agent can spend funds independently, there must be safeguards around limits, approvals, fraud detection and recovery mechanisms.

Multi-party custody models could allow human oversight, programmable spending rules and distributed authorization structures that reduce the risks of fully autonomous financial control.

Crypto rails are increasingly seen as machine-native infrastructure

The broader idea behind the discussion is that crypto may ultimately succeed not because humans love wallets, but because programmable software systems do. Blockchain networks already support internet-native value transfer, programmable settlement and API-based financial coordination.

For AI agents operating globally across digital platforms, those features may fit naturally with how autonomous systems interact online. That does not guarantee mass adoption, but it explains why AI and crypto narratives are increasingly converging.

Why this matters

This matters because it reframes crypto adoption in a completely different way. Instead of asking whether ordinary consumers want to use blockchain technology directly, some companies are beginning to ask whether AI agents might become crypto’s real long-term users.

If autonomous commerce becomes widespread, the infrastructure supporting it may need to be programmable, open and globally interoperable. Those are all areas where crypto systems already specialize.

The clean takeaway

Executives from PayPal and Google Cloud believe future AI-driven commerce may rely heavily on crypto-style infrastructure. Open payment protocols, programmable custody and machine-readable commerce systems could become essential if autonomous AI agents eventually start buying, selling and transacting online at scale.