Новини

Sapiom Raises $15M to Help AI Agents Buy Their Own Tech Tools

· TechCrunch

The rise of no-code AI development platforms like Lovable has enabled non-technical users to build custom applications through simple text descriptions. However, these "vibe coding" solutions face a critical limitation: connecting applications to external services like SMS providers, payment processors, and cloud infrastructure requires complex backend integration that often stumps non-technical creators. Sapiom, a San Francisco startup founded by former Shopify payments director Ilan Zerbib, is addressing this infrastructure gap with an innovative approach. The company has raised $15 million in seed funding led by Accel to build what it calls "the financial layer" for AI agents—a payment system that enables AI to automatically purchase and access the software, APIs, data, and compute resources it needs. The core problem Sapiom solves is authentication and micro-payments for AI agent interactions. Every time an AI agent connects to external tools like Twilio for SMS messaging, it requires proper authentication and payment processing. Currently, this process is manual and complex, requiring developers to sign up for services, add payment methods, and integrate API keys. Sapiom automates this entire workflow, allowing AI agents to make independent purchasing decisions without human intervention. "In the future, apps are going to consume services which require payments. Right now, there's no easy way for agents to actually access all of that," explained Amit Kumar, Accel partner who led the investment round. Kumar's thesis centers on the enterprise focus rather than consumer applications, believing this B2B approach is essential for making AI agents truly autonomous. The funding round attracted notable participation from Okta Ventures, Gradient Ventures, Array Ventures, Menlo Ventures, Anthropic, and Coinbase Ventures, signaling strong industry confidence in Sapiom's vision. The diverse investor base reflects the cross-industry implications of AI agent autonomy. Sapiom's technology works as a pass-through payment system. For example, when someone builds an SMS-enabled app using a vibe-coding platform, they won't need to manually configure Twilio integration. Instead, Sapiom handles the service provisioning in the background, and the user gets charged for Twilio services as a pass-through fee by their chosen development platform. While currently focused on B2B solutions, Sapiom's technology has broader implications for consumer AI agents. The company envisions a future where personal AI assistants can independently handle transactions like ordering rideshares or making purchases. However, Zerbib remains pragmatic about consumer adoption, noting that AI won't necessarily drive increased consumer spending—hence the strategic focus on business applications. The timing appears optimal as the AI agent ecosystem matures. With prompt-to-code platforms gaining traction and enterprises increasingly deploying AI agents for various tasks, the need for seamless service integration becomes critical. Sapiom's infrastructure could become the backbone enabling true AI agent autonomy, removing the technical barriers that currently limit non-technical creators and enterprise AI deployments.

The democratization of app development through AI-powered "vibe coding" platforms has opened new possibilities for non-technical creators, but a critical infrastructure gap remains. While tools like Lovable enable users to transform plain-language descriptions into working applications, the challenge of integrating these prototypes with external services—from SMS providers to payment processors—often proves insurmountable for those without coding backgrounds.

Sapiom, a San Francisco-based startup founded by Ilan Zerbib, former director of engineering for payments at Shopify, is tackling this fundamental limitation with an ambitious vision: creating a financial layer that enables AI agents to autonomously purchase and access the software, APIs, data, and compute resources they need to function.

The company has secured $15 million in seed funding led by Accel, with participation from a notable roster of investors including Okta Ventures, Gradient Ventures, Array Ventures, Menlo Ventures, Anthropic, and Coinbase Ventures. This diverse investor base reflects the cross-industry implications of Sapiom's technology and the growing recognition that AI agent autonomy requires sophisticated financial infrastructure.

The core challenge Sapiom addresses is both technical and financial. Every interaction between an AI agent and external services requires authentication and micro-payments. When an AI agent needs to send an SMS through Twilio, spin up AWS compute resources, or process a Stripe payment, it must navigate complex authentication protocols and payment mechanisms. Currently, this process is entirely manual, requiring developers to sign up for services, configure payment methods, and integrate API keys—a significant barrier for non-technical users.

"If you really think about it, every API call is a payment. Every time you send a text message, it's a payment. Every time you spin up a server for AWS, it's a payment," explained Amit Kumar, the Accel partner leading the investment. Kumar's extensive research into AI payments startups convinced him that Zerbib's enterprise-focused approach represents the most viable path to enabling true AI agent autonomy.

Sapiom's solution functions as an intelligent intermediary, handling service provisioning and payments automatically. In practice, this means that someone building an SMS-enabled app through a vibe-coding platform won't need to manually configure Twilio integration. Instead, Sapiom manages the entire backend process, with users receiving pass-through charges for consumed services via their chosen development platform.

The implications extend far beyond simplifying app development. As AI agents become more sophisticated and autonomous, they'll need the ability to independently access and purchase services to complete their tasks. Sapiom's infrastructure could enable scenarios where AI agents automatically scale compute resources based on demand, purchase additional API credits when needed, or access specialized data services to enhance their capabilities.

While Sapiom currently focuses on B2B applications, the technology has clear consumer implications. The company envisions a future where personal AI assistants can independently handle transactions—ordering rideshares, making purchases, or booking services—without requiring human intervention for each payment authorization. However, Zerbib maintains a pragmatic perspective on consumer adoption, noting that AI won't necessarily drive increased consumer spending, which reinforces the strategic focus on business applications where the value proposition is clearer.

The timing of Sapiom's emergence appears optimal. The vibe-coding ecosystem is gaining momentum, with platforms like Lovable and Bolt making app development accessible to non-technical users. Simultaneously, enterprises are increasingly deploying AI agents for various operational tasks, creating demand for seamless service integration capabilities.

The $15 million seed round positions Sapiom to build the foundational infrastructure that could become essential as AI agents evolve from simple chatbots to autonomous digital workers. By solving the financial layer challenge, Sapiom is addressing a critical bottleneck that currently limits the potential of AI agent technology.

As the AI ecosystem continues to mature, the ability for agents to independently access and purchase services may become as fundamental as internet connectivity itself. Sapiom's vision of frictionless AI-to-service transactions could unlock new possibilities for both enterprise automation and consumer AI applications, making the promise of truly autonomous AI agents a practical reality.

Посилання в статті