Coinbase is testing AI agents inside Slack that are modelled on two of its own founders, as CEO Brian Armstrong says the company will eventually employ more AI agents than human workers.
The initiative, According to CEO Brian Armstrong, signals a decisive pivot toward automation in one of the crypto industry’s most influential firms.
The Coinbase AI agents are designed to support employees in handling routine and strategic tasks, effectively acting as digital coworkers embedded within daily communication tools. Armstrong said the company has already deployed two such agents, each tailored to different operational roles, with early results pointing to improved efficiency and faster decision-making.
This move aligns with Coinbase’s long-term ambition to embed artificial intelligence deeply into its corporate structure. According to Armstrong, Coinbase AI agents could soon become a dominant force within the organization. “We will have more agents than human employees at some point soon,” he noted, underscoring the scale of the company’s vision.
The introduction of Coinbase AI agents also opens the door for employees to build custom agents for their teams, further decentralizing innovation within the company. This approach could transform how internal tools are developed, shifting from centralized engineering to user-driven AI customization.
AI Models Inspired by Leadership Drive Strategic Thinking
The first generation of Coinbase AI agents reflects a unique design philosophy: modeling artificial intelligence on real-life leadership figures. One agent, named Fred, is based on co-founder Fred Ehrsam and functions as a strategic advisor. It helps employees prioritize tasks and provides feedback akin to executive-level guidance.
The second of the Coinbase AI agents, Balaji, draws inspiration from former CTO Balaji Srinivasan. This agent is engineered to foster creativity, challenge assumptions, and encourage unconventional thinking. By simulating different leadership styles, Coinbase AI agents aim to broaden how employees approach problem-solving.
This dual-agent system highlights a broader trend in enterprise AI: the move from purely functional tools to systems that replicate human reasoning patterns. Coinbase AI agents are not just automating tasks—they are attempting to augment cognitive processes within teams.
A Coinbase spokesperson described the goal as enabling employees to “see problems from multiple perspectives,” a capability that could enhance both strategic planning and product development. The company believes Coinbase AI agents can act as force multipliers, especially in high-stakes decision environments.
Building an AI-Native Workforce
The rollout of Coinbase AI agents is part of a larger transformation aimed at making the company “AI-native.” With a workforce exceeding 4,000 employees, Coinbase is investing heavily in training staff to work alongside AI systems across coding, analytics, and internal communications.
Armstrong has previously stated that more than half of Coinbase’s codebase could eventually be generated by AI. In this context, Coinbase AI agents serve as both tools and training mechanisms, helping employees adapt to a future where machine-assisted development is the norm.
The company’s technological push includes the introduction of the x402 protocol in 2025, which facilitates payments for AI agents using both cryptocurrency and traditional financial rails. This infrastructure is critical for enabling Coinbase AI agents to operate autonomously in transactional environments.
By combining AI capabilities with crypto-native payment systems, Coinbase is positioning its agents not just as internal assistants but as participants in a broader digital economy. This could allow Coinbase AI agents to execute tasks such as paying for services, accessing APIs, or even negotiating transactions without human intervention.
Coinbase AI Agents Reflect Broader Industry Shift
The rise of Coinbase AI agents mirrors a growing consensus across the crypto sector: autonomous agents are poised to become key actors in online ecosystems. Armstrong has emphasized that AI-driven transactions could soon surpass those conducted by humans, a view echoed by other industry leaders.
Coinbase AI agents are particularly well-suited to this environment due to crypto’s inherent advantages—speed, programmability, and global accessibility. These features make blockchain-based systems an ideal backbone for machine-to-machine commerce.
Industry analysts suggest that Coinbase AI agents could eventually handle a wide range of activities, from executing trades to managing subscriptions and coordinating logistics. In such a scenario, human oversight would shift from execution to governance and strategy.
However, the rapid deployment of Coinbase AI agents also raises questions about workforce displacement and ethical considerations. While the company frames its AI push as a productivity enhancement, the implication of fewer human workers is difficult to ignore.
Still, Coinbase maintains that its focus is on augmentation rather than replacement. By equipping employees with Coinbase AI agents, the company aims to increase output while enabling staff to focus on higher-value tasks.
As the experiment unfolds, Coinbase AI agents could set a precedent for how technology firms integrate artificial intelligence at scale. If successful, the model may be replicated across the crypto industry and beyond, marking a significant step toward an automated, AI-driven economy.
