Goldman Sachs is developing AI agents with Anthropic to automate key back-office functions. The partnership shows how major financial institutions are moving beyond basic AI experiments into practical automation.
The bank has worked with embedded Anthropic engineers for six months. They are building autonomous systems for accounting and compliance work.
Marco Argenti, Goldman’s chief information officer, said the agents will reduce the time needed for essential tasks. The systems will handle trade accounting and client vetting processes.
The bank started testing AI coding tools last year with a product called Devin. Goldman then discovered Anthropic’s Claude model could work in other areas.
The AI agents will work on jobs that combine data processing with rule application. These include compliance checks and transaction reconciliation.
Goldman executives found Claude capable of handling complex, rules-based work. The model can parse large amounts of documents while applying judgment.
Argenti said clients will be onboarded faster once the agents launch. Trade issues will also be resolved more quickly.
The bank could expand AI agents to other tasks. Future applications might include employee surveillance and creating investment banking presentations.
Goldman employs thousands in compliance and accounting roles where AI will operate. Argenti said it was too early to predict job losses.
The bank may reduce its use of third-party providers as AI improves. For now, Goldman views the technology as adding capacity rather than replacing workers.
Anthropic is finalizing a funding round exceeding $20 billion. The deal would value the company at about $350 billion.
The startup initially aimed to raise $10 billion. Strong investor demand pushed the target higher.
This round would nearly double Anthropic’s previous valuation. The company raised $13 billion just months ago.
Coatue Management, Singapore’s GIC, and Iconiq Capital will each invest over $1 billion. Nvidia and Microsoft could contribute as much as $15 billion combined.
Other investors include Altimeter Capital Management, Sequoia Capital, Lightspeed Venture Partners, and Menlo Ventures. The wide participation shows continued investor appetite for AI leaders.
Anthropic’s revenue run rate crossed $9 billion last summer. The company’s coding tools have gained traction with developers and large firms.
The startup recently released a model for automating work tasks. The announcement triggered a selloff in software stocks as investors assessed AI’s impact on existing tools.
The funding highlights the high cost of building advanced AI. Companies spend heavily on chips, data centers, and specialized talent.
Anthropic ran a Super Bowl ad stating ads would not appear in Claude. The commercial warned that other AI chat services might show advertisements.
Co-founder Daniela Amodei called ads in AI chats “exploitative.” She said users share personal information that shouldn’t be monetized through advertising.
OpenAI CEO Sam Altman responded on X after the ad aired. He said OpenAI would not run ads in the way Anthropic depicted.
OpenAI is reportedly in talks to raise up to $100 billion. Both companies have explored possible IPO paths this year.
The post How Goldman Sachs Plans to Automate Thousands of Banking Jobs with Claude AI appeared first on CoinCentral.


