Many experts share Bilotta’s AI agent view, including Charles Hoskinson, founder and CEO of Cardano’s Input Output, who said that by 2035 they will become more relevant than humans.
The macro figures support his position. Data from the US International Trade Administration via Trade.gov shows that business-to-business (B2B) e-commerce across the wider Asia-Pacific region is expanding at an annual clip of 15%, with market values expected to climb above US$28.9 trillion by the end of this year.
Yet, despite the explosive growth, the plumbing underneath remains broken. The problem is fundamentally one of legacy infrastructure and compliance, Bilotta noted.
Global financial rules, banking protocols and identity verification checks were built exclusively for humans. An autonomous AI software agent cannot pass a standard compliance check, he said, or perform a payment loop unless a human manually intervenes to clear the transaction.
To bridge this structural gap, the industry requires a compatible backend middleware that acts as a universal interpreter. Bilotta explained that by dropping an anthropic standard Model Context Protocol (MCP) server directly into the payments infrastructure, software agents can programmatically navigate compliance, pull real-time FX rates and settle cross-border transactions without human steps.
While institutional gatekeepers like Stripe and Mastercard have spent billions acquiring fiat-to-crypto APIs to secure traditional corporate taxes, the automated machine-to-machine economy across new corridors remains vastly undervalued.
