Mounir Mouawad is a product entrepreneur, co-founder and former chief executive of Portia AI, and a director of AgentBay and AgentBay Technologies. After building payments and commerce products at Amazon, Google, and Stripe, he helped develop infrastructure for AI agents that must operate securely, explain their actions, and remain accountable to human operators.
From payments infrastructure to AI startups
Mouawad studied computer engineering at the American University of Beirut and earned an MBA from London Business School. At Amazon, he launched the United Kingdom deals program supporting events including Black Friday and Prime Day. He subsequently expanded Google Pay across 30 markets in Europe, the Middle East, and Africa, served on Google Payments’ United Kingdom board, and led banking-as-a-service in EMEA at Stripe.
At Stripe, he met Emma Burrows, who became his Portia AI co-founder. Incorporated in 2024, Portia developed an open-source agent framework and cloud platform addressing the control requirements of regulated businesses; General Catalyst backed the company in 2025. Mouawad advocated predictable, auditable AI agents that expose execution plans, report progress, authenticate securely, and request human intervention when necessary, principles he outlined in a public introduction to Portia.
Mouawad left Portia’s board in September 2025. He became a director of AgentBay in March 2026 and AgentBay Technologies in July 2026.
What makes agents practical
Browser automation needs multimodal perception. While testing an Amazon shopping agent, Mouawad encountered brittle page-specific instructions, oversized HTML inputs, login barriers, and anti-bot defenses. His analysis of browser agents argues that combining visual and textual interpretation makes navigation less dependent on fragile webpage parsing.
Delegated authentication and durable human oversight. Mouawad favors delegated authentication for AI agents, enabling websites to distinguish authorized software from malicious automation. He also argues that agents should pause for logins, payment approvals, and other consequential decisions, preserve their execution state, and resume when human authorization arrives.
Customer discovery precedes confident roadmaps. Mouawad champions hypothesis-driven product discovery because potential customers often recognize AI’s promise before identifying a useful workflow. His AI Engineer founder presentation describes testing focused product bets, adapting quickly to standards such as the Model Context Protocol, and earning distribution through developer communities, partnerships, credible advocates, and integration documentation.
