Via cryptorank.io
The crypto custodian's new AI tool lets users query account data and platform features in plain English, but deliberately cannot execute transactions.
BitGo just rolled out an AI assistant that lets institutional clients ask questions about their accounts the way they’d ask a colleague. BitGo Assist, launched on August 12, is a natural language search agent embedded directly into the company’s platform, designed to surface answers about workflows, account data, and platform features without requiring users to click through menus or read documentation.
The key design choice: it deliberately cannot execute, approve, or sign any transactions.
What BitGo Assist actually does #
The tool lets clients type queries in plain language to navigate BitGo’s institutional platform, which spans custody, trading, staking, and settlement services. Need to understand a complex digital asset concept? Ask Assist. Want to pull up your account balances? Type the question instead of hunting through dashboards.
BitGo Assist works within the user’s authenticated session, meaning it only surfaces information the logged-in user already has permission to see. It uses data tokenization to protect sensitive details like account numbers and wallet addresses, adding a layer of security on top of the session-based access controls.
The agent can also initiate steps required for various inquiries, directing users to the right part of the platform when they need to take action.
Part of a broader AI strategy #
BitGo Assist represents the second phase of the company’s AI integration roadmap. The first phase arrived in March 2026 with the launch of BitGo’s MCP Server, which laid groundwork for machine-readable platform interactions.
The company has framed its AI efforts around a concept it calls “client autonomy,” the idea that institutional users should be able to self-serve complex operations without needing to contact support teams or wait for account managers.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our