FinTech | Payments | Banking | Innovation | Leadership
Are #AI agent #payments the next big thing? Is science fiction becoming reality? And how much is the hype vs the potential? Let’s take a look. What are AI agent payments? — When we talk about agent payments, we mean letting AI agents handle payments on our behalf. — AI agents are AI #software programs capable of autonomously performing tasks based on broadly defined, open-ended goals. They use tools, make decisions, and take actions to achieve their goals. Why are AI agent payments important? — Because it can be a directional, cross-industry pivot, radically changing #strategy, business models and monetization. — Because it might represent the build-up of a next-gen payments’ infrastructure layer that will work on top (but not replace) of existing platforms and methods. Who are the players? 1. Stripe: — It’s the latest joiner (and by far the biggest player) by releasing a few days ago the Stripe Agent Toolkit, a software development kit (SDK) enabling Payment and Billing Capabilities for AI Agents. — Killer use case: the combination with Stripe Issuing where agents create single-use virtual credit cards (spend controls as safeguards) to pay for products. Automating of the purchase flow combined with embedded financial services! 2. Coinbase: — Multi-party computing (MPC) wallet launched by Coinbase Developer Platform (CDP). It can be integrated with AI Agents to enable them to pay autonomously. — Self-hosted crypto wallets to provide support for AI Agents. AI to AI crypto transactions. 3. Skyfire: — A payment network specifically for AI agents. — Agent funding via traditional rails and with USDC (USD stablecoin). — Bet on the growing agent network and interoperability with other frameworks. 4. Protegee: — A startup focusing on voice payments via AI voice agents accepting payments over the phone. — AI call centers are a typical use case with Protegee agents collecting booking details, processing transactions, and returning back payment confirmation. 5. Payman: — Pre-seed startup invested by Visa. — AI to Human payments platform. — A human marketplace for AI: building a marketplace for AI to find humans to get assistance in completing tasks. — Multiple Payment Rails: fiat, crypto, bank account. My main takeaways: 1. This a race to capitalize on the new AI agent economy, which is why interest is huge with over $1 billion invested in 2024 alone. 2. It’s an infrastructure game taking place across several layers. 3. It’s a matter of time before more big players join the race (directly or indirectly). 4. Crypto is emerging as a promising rail, mainly because of facilitating agent-to-agent transactions without identity verification. 5. Main challenges across 2 directions: a) Regulation (i.e. is KYC or AML needed and how do you perform them on AI agents?) b) Controls and safeguards for bots managing our personal and business accounts. Opinions: my own
Discussions around AI payments and the DEPIN framework remind me of the early IoT days, when we envisioned a 5G-powered future where every device would not only communicate seamlessly but also generate income for its owner. I’d love to believe that one day AI agents and digital twins will become so integrated into our lives that we can sip cocktails while they do all the heavy lifting. But, knowing how big corporations operate, I remain skeptical for now.
Solutions looks great but with the challenges of brick-and-mortar banking systems this is difficult to achieve due to compliance and IT policies. Plus, trust is a big issue too for retail users. With so many AI systems generating deepfakes could cannibalize such niche solutions if the organization lose sight of security leading to financial loss. Finally, without CBDC infra these solutions are “good to have” instead of “must have”.
The regulatory challenges around KYC for AI agents are crucial. Having built payment systems across markets, I can say this will be one of the most interesting puzzles to solve.
I dont know how many of you have been to China, AI payemente is implemented there. The use case: show your face at rail station Tap in/tap out. smile and transaction is exectued. The Machine learning or Agents recognises the face, I am also concenerd about the safety financial transactions at e commerece level, at pay points, and at restuarants, mainly. Where i am not concerned is internet of things, the system (IOT) already knows who you are, where do you go what tasks will you perform each day , How many payaments do u do each day. Margin of error (materiality ) is embendee the the system by apply machine learnimg. The freqenency is learned. Now since the ai agents have been arrived they will Come with the ability to think, raise objection where payment is not to be made. agent do keep the ability to learn about the volnuerabilities in any payment system basically from the backened data and payment infrastructr , ( be it at the conventional clouds or on any blockchain). The system colnuerabilities arises when these agents volnuerably transfer the colnuerable data to by pass the system. Long way to go on AI PAYMENT and internetof thingsCant say much then these basics.
This is a fascinating exploration of the potential for AI agents in payment processing, Panagiotis. The transition from traditional models to AI-driven systems could indeed redefine our approach to transactions and financial interactions. I'm particularly intrigued by the implications for regulation and the need for robust safeguards. Your insights are timely and critical as we navigate this transformative landscape.
AI agent payments will revolutionise multiple sectors. One interesting application could be in the travel industry, where travel agents could leverage AI agents to pay suppliers more efficiently with single use virtual cards or indeed crypto. This could streamline the process and ensure more timely payments. Watching how this develops will be fascinating!
Fascinating dive into the evolving AI agent payments space! Combining AI, fintech, and crypto is a game-changer for automation and financial infrastructure. Stripe's virtual card issuing, Coinbase's crypto-first vision, and the rise of agent interoperability point to an exciting future. Regulation & safeguards will be critical, but the potential for autonomous, AI-to-AI commerce is mind-blowing!
It sounds wonderful. Don’t think this will happen any time soon (15 years) as the governance and controls in the financial sector are too heavy.
A couple of problem with this sequence: 1: most CBDCs, especially in advanced economies won't be using blockchain. 2: Smart contract aren't legal contracts and aren't really smart in being able to make judgment, they are preprogrammed. 3: AI cannot be proven correct, proven to behave consistent or proven to not be undermined i.e. to drain an account Then it ignores the operational costs of running blockchains and AI. As it stands payments cost society already too much. The challenge for future payments is to be digital. cheaper with better control by the owner. That's mot what this shows.