Google's AP2 Explained: Why FastSet Is the Natural Foundation

Google just announced the Agent Payments Protocol (AP2), a new standard for how AI agents interact and handle payments on our behalf. 

Having AI agents transact on behalf of users challenges the assumption that a human is always the one pressing “buy.” This shift raises key issues of authorization (was the agent truly allowed?), authenticity (does the request reflect the user’s intent?), and accountability (who’s responsible if something goes wrong).

AP2 tackles these challenges by introducing an open standard for secure, compliant transactions, built around mandates: cryptographically signed digital instructions that record either a user’s intent (e.g., “buy this if under $200”) through intent mandates, or the resulting cart details (e.g., “these items at these prices”) through cart mandates.

This is an important step forward. It gives agents a structured, auditable way to transact securely. 

Reading about AP2, it’s clear: its design philosophy mirrors what we’ve been building with FastSet all along. AP2’s mandates can be seen as specialized cases of FastSet claims. An intent mandate directly corresponds to an intent claim in FastSet, cryptographically signed by the user, with the signature serving as verifiable proof of correctness (what AP2 refers to as verifiable credentials). Similarly, a cart mandate maps to a specialized claim in FastSet that defines cart items and their properties, cryptographically signed by the agent as proof. Put simply, AP2 represents a domain-specific specialization of FastSet’s broader claims architecture for AI agent interactions, making FastSet the natural foundation for deploying AP2 at scale.

In fact, with FastSet, AP2 can be endowed with powerful extensions:

• Uniform treatment of interactions and payments. User intents, agent actions, and payments are all expressed as claims in the same formalism. This unification eliminates silos between different transaction types, making the entire lifecycle, from intent to execution to settlement, auditable, composable, and verifiable under a single framework.
• Flexible verifiability. Agent actions and payments don’t need to rely solely on cryptographic signatures. FastSet supports richer forms of verifiable proofs. For instance, a high-value purchase could require an independently verifiable certificate from a network of specialized e-commerce verifiers to ensure the cart truly reflects the original intent. Conversely, private purchases could be backed by zero-knowledge proof to protect sensitive details. FastSet gives developers the flexibility to choose the right proof system for each use case.
• Decentralized accountability. Instead of depending on a small set of intermediaries, claims are validated by independent validators and settled with nearfinality on FastSet. This creates a transparent, tamper-proof trail that guarantees accountability without central choke points.
• Agent-to-agent at scale. Millions of agents running 24/7 will generate astronomical volumes of claims. FastSet’s breakthrough architecture — massively parallel processing, sub-100 ms finality, and practically unbounded throughput — is purpose-built to handle this load at global scale.
• Generalization beyond payments. While AP2 focuses on payments, FastSet serves as a universal substrate for verifiable settlement — spanning commerce, data exchange, supply chains, digital assets, and even AI-to-AI coordination.

An implementation of AP2 on FastSet can power a future where agents transact seamlessly, securely, and at scale, with trust minimized and accountability guaranteed.

At Pi Squared, we’re excited to see this convergence. It validates our vision and highlights why a fast, verifiable settlement layer is essential for the agentic economy.