No reading or writing, just "acting": that's how the new internet will work, where AI will do the heavy lifting for you.

Franklin Templeton is analyzing the transition to an Internet Agent, where protocols such as A2A, MCP, and AP2 enable AI to perform autonomous financial tasks and transactions.

The evolution of the digital network has reached a new turning point that promises to redefine the digital economy. According to a recent analysis published by asset manager Franklin Templeton, We are witnessing the birth of the "Internet Agent", a phase where the web ceases to be a merely readable or writable space to become "actionable". 

For Franklin Templeton, the premise is clear: if the last decade was defined by platforms where humans created content, the next will be dominated by standalone software capable of reasoning, negotiating and, crucially, carrying out economic transactions without constant human intervention.

The firm emphasized that this paradigm shift is not based on abstract futurology, but on the convergence of new infrastructure technologies that are being quietly adopted by tech giants. Its report highlights the capacity of the Large Language Models (LLM) Planning complex tasks has gone from being a novelty to becoming the operational foundation of a new digital workforce. However, it was also noted that for these agents to be truly useful, they need to move beyond the chat window and interact with the real world—a leap that requires new communication and payment standards that are only just beginning to be implemented.

An invisible infrastructure for the machine economy

Franklin Templeton's analysis details that, just as the HTTP and TCP/IP protocols were the foundation of the early web, the Internet Agent is being built on three new technical standards designed to resolve current fragmentation. First, there is the Agent to Agent protocol, or A2AUntil now, artificial intelligences operated in closed silos; A2A proposes a shared language that allows a purchasing agent, for example, to communicate fluidly with a market research agent, coordinating complex tasks across different organizational platforms.

For this collaboration to be effective, the agents need access to information. This is where the Model Context Protocol (MCP), described in the Valid identity document like a "universal plug" for data. 

Instead of building fragile, custom integrations for each database or application, MCP allows AI systems to connect in a standardized way to enterprise tools and files. According to the report, companies like Anthropic and Microsoft are already integrating these standards, suggesting that the technical barrier to AI accessing corporate information is about to collapse, enabling operations that go far beyond simply answering questions.

Programmable money as an engine of AI autonomy

Perhaps the most relevant aspect for the financial ecosystem is how these agents will manage value. Franklin Templeton's view is pragmatic: an agent who cannot pay is merely a glorified assistant. 

To solve this, the report introduces the AP2 protocol (Agent Payments Protocol) and its crypto-native extension, x402. AP2 functions as a governance layer that manages "mandates"—cryptographically signed instructions where the human user defines spending limits and purchase conditions. This allows an AI to have the authority to execute payments, but always within a verifiable framework of trust.

However, traditional bank rails, with their cut-off times and latency, prove inefficient for a machine economy operating 24/7. This is where the analysis underscores the critical role of the cryptocurrencies, and stablecoins. 

Through the standard x402 (Crypto Rail Extension)Instant and programmable settlement is enabled. This opens the door to microtransactions that would be impossible with conventional credit cards, allowing agents to pay pennies for access to APIs, data, or real-time services. The participation of players like Circle in the pilot tests of these protocols reinforces the idea that blockchain technology will find its "ultimate use case" as the financial nervous system of artificial intelligence.

Certificate

Blockchain Course

Basic level

Take this course where we explain blockchain in a clear, simple and concise way so that you have a very clear idea of ​​what this new technology consists of.

FREE 25 lessons

Start now

The next frontier: digital infrastructure will redefine the economy of the future

Franklin Templeton observes with caution, but with conviction, the immediate path ahead for the technology sector. It is projected that, in the coming years, certain technical specifications will solidify, paving the way for widespread adoption of these technologies toward the end of the decade.

According to the investment manager, what's relevant for investors and market participants is that economic value will likely shift. It will no longer be just about who owns the smartest AI model, but about who controls the trust infrastructure: digital identity issuers, payment mandate validators, and stablecoin liquidity providers.

The report indicates that in the near future, true economic value will shift toward these key infrastructure points. This means that software capable of automating end-to-end business processes will transform the so-called digital bureaucracy—which includes identity verification and automated regulatory compliance—into a strategic and profitable asset. In this vision, blockchain technology and artificial intelligence will cease to advance along parallel paths and will instead integrate into a homogeneous system, revolutionizing how capital flows through global networks.