Coinbase has begun testing artificial intelligence agents that integrate directly into internal communication tools, a move that signals a deeper shift toward automation inside the company’s workforce. The initiative places AI agents in platforms such as Slack and email, where they appear and respond in a way that resembles human colleagues.
CEO Brian Armstrong confirmed the rollout in a post on X, where he described the agents as teammates that can participate in daily workflows.
"Coinbase is testing AI agents that show up in Slack/email at work, just like any human teammate. To start, we’re shipping two, which are modeled after legendary former Coinbase employees, Fred Ersham and Balaji Srinivasan."
‘Fred’ and ‘Balaji’ introduced as internal AI teammates
The first two agents reflect distinct roles tied to their real-world inspirations. Fred Ehrsam serves as the model for "Fred," which functions as a strategic executive agent. The tool provides high-level feedback and helps refine internal documents, strategy drafts, and planning frameworks.
The second agent draws from Balaji Srinivasan, a former chief technology officer known for his work in crypto and as the author of The Network State: How to Start a New Country. Labeled internally as an "agent of chaos and creativity," the Balaji model challenges assumptions and supports long-term thinking. It aims to push employees toward unconventional approaches and new ideas.
Travis Bloom, an engineer at Coinbase, shared an early experience with the system. He said that a conversation with the Balaji agent helped clarify a new concept, noting that it "helped crystallize" his vision.
Expansion plans and internal experimentation
Armstrong described the current rollout as an early phase. He indicated that Coinbase plans to expand the number of agents and lower the barrier for employees to create their own versions.
"Soon, it will be easy for any employee to spin up a new agent for themselves or their team," he said.
He added that the number of AI agents could eventually surpass the number of human employees at the company.
The CEO also suggested a shift away from framing these systems as replicas of real individuals.
"I think these employee agents should have their own name actually. Not a ‘digital twin’ of someone else. So that is next step," he wrote in a separate post.
AI integration aligns with broader company strategy
The experiment connects with Coinbase’s wider push into AI-driven infrastructure. The company has already introduced "Agentic Wallets," which allow AI systems to hold funds, execute transactions, and interact with blockchain networks without direct human control. These wallets operate alongside Coinbase’s x402 protocol, which focuses on enabling autonomous payments across crypto and fiat systems.
Armstrong has previously outlined ambitions to automate a large portion of internal workflows. He stated in earlier remarks that he wants more than 50% of Coinbase’s codebase to be generated by AI. The company has also emphasized transforming its workforce into "AI-natives," a shift that affects how employees interact with tools, data, and decision-making processes.
Industry context and emerging concerns
Coinbase’s approach arrives at a time when AI adoption accelerates across both technology and crypto sectors. Companies have increased investment in automation while reducing reliance on traditional roles. The integration of AI agents into communication channels marks a shift from passive tools toward active participants in organizational structures.
However, the model raises questions about accountability and oversight. AI agents do not represent real individuals, even when trained on their communication styles or strategic thinking. This creates uncertainty around responsibility for decisions or recommendations generated by such systems.
At the same time, some leaders in the crypto sector have pointed to a future where AI agents play a central role in digital economies. Armstrong has stated that AI agents could outnumber humans in online transactions. Jeremy Allaire has made similar remarks, predicting that billions of AI agents could transact onchain within a few years. Changpeng Zhao has also argued that cryptocurrency could serve as the native payment layer for these systems.
A shift toward AI-native organizations
Coinbase’s internal testing reflects a broader transition toward AI-native companies, where software agents assist with strategy, execution, and communication. The use of personalities modeled after former executives adds a layer of familiarity, but the company has already signaled that future iterations will move beyond direct imitation.
The early feedback from employees highlights practical use cases in brainstorming, planning, and decision support. However, the long-term implications remain unclear, particularly in areas such as governance and trust.
As the company continues to expand the program, its experiment could influence how other firms structure their teams, blending human and AI contributions inside a shared digital workspace.
Emmanuel Ekokotu
Disclaimer: All materials on this site are for informational purposes only. None of the material should be interpreted as investment advice. Please note that, despite the nature of much of the material created and hosted on this website, HODL FM operates as a media and informational platform, not a provider of financial advisory services. The opinions of authors and other contributors are their own and should not be taken as financial advice. If you require advice, HODL FM strongly recommends contacting a qualified industry professional.
