Goldman Sachs has taken a decisive leap into the future of software development, announcing the deployment of an autonomous artificial intelligence engineer known as Devin, developed by AI startup Cognition, in a move that could reshape how Wall Street firms approach technical operations.
Goldman’s Chief Information Officer Marco Argenti confirmed to CNBC this week that the bank is now actively testing Devin, with plans to initially roll out hundreds of these AI agents to support its 12,000 human engineers—and possibly expand the use to thousands, depending on its success.
“We’re going to start augmenting our workforce with Devin, which is going to be like our new employee who’s going to start doing stuff on the behalf of our developers,” Argenti said.
From Assistants to Agents
Unlike traditional AI assistants that help summarize emails or search databases, Devin represents a new wave of agentic AI—programs that don’t just assist but actually perform multi-step software engineering tasks independently. The AI agent operates as a full-stack developer, capable of writing, debugging, testing, and deploying code with minimal human intervention.
Demo videos from Cognition last year showcased Devin autonomously handling assignments that would typically require a team of human engineers. These include upgrading internal software systems, migrating legacy code, and developing entire applications from scratch.
From Buzz to Buy-In
Just over a year ago, firms like JPMorgan Chase and Morgan Stanley began cautiously introducing OpenAI-powered cognitive tools across their operations. Today, Goldman’s embrace of Devin reflects how fast the frontier has shifted from experimentation to integration.
At Goldman Sachs, Argenti noted that Devin’s productivity potential far exceeds the earlier generation of AI tools, claiming it could boost productivity by up to three or four times compared to previous systems.
“Initially, we will have hundreds of Devins, and that might go into the thousands,” he added, signaling that the firm sees this as a long-term investment in operational efficiency.
Goldman plans to deploy Devin to take over repetitive but essential coding tasks—like updating software to newer programming languages or refactoring legacy systems—allowing human engineers to focus on high-value, strategic innovation.
Wall Street Joins Silicon Valley in the AI Code Race
The adoption of AI coders isn’t unique to Goldman. Big Tech firms like Microsoft and Alphabet have reported that AI tools are already responsible for writing 30% of all code on major projects. At Salesforce, CEO Marc Benioff recently claimed AI now performs up to half of the company’s total work.
But Goldman Sachs’ deployment of Devin is especially symbolic: Wall Street’s most elite investment bank is no longer just investing in AI—it’s hiring it.
Supervised, But Autonomous
Despite its autonomous capabilities, Argenti emphasized that Devin will still operate under human oversight to ensure accuracy and avoid unintended errors.
“We see Devin not as a replacement, but as a multiplier,” said Argenti. “Our developers will now be managing AI counterparts, not competing with them.”
However, the expansion of AI “employees” raises broader questions about the future of work, especially for junior developers and operations staff whose tasks could increasingly be automated.
The move underscores a broader trend in finance and tech: AI is not just a tool—it’s becoming a co-worker.
With Devin, Goldman Sachs isn’t just testing the waters, it’s jumping headfirst into a future where Wall Street coders may be just as likely to review machine-generated code as they are to write it.
If successful, Devin’s rollout could inspire similar deployments across other major financial institutions, hastening the rise of AI-powered enterprise teams—and redefining what it means to “hire” in the age of intelligent machines.