Bolt-on AI (competitors)
AI layer added to a 2010 system. Permissions inherit from the original design. Evidence and audit are secondary. Tool governance is limited by legacy boundaries.
AI-native MCP
QuantJourney exposes investment workflows to AI agents through controlled tool calls — portfolio state queries, thesis retrieval, regime signal access, Decision Packet preparation. Every query is permissioned, every response cites its source, every action leaves an audit record.
Why AI-native is different from AI bolt-on
Aladdin Copilot. eFront Copilot. Addepar Addison. All are generative AI experiences built on top of platforms designed before AI was a requirement. The architecture matters. Bolt-on AI operates within the permissions and data boundaries of a system designed for a different purpose. Evidence, audit trails and tool governance are afterthoughts. QJ is AI-native by design — MCP-based tool calls, permissioned data access, evidence-cited responses, every agent action auditable.
AI layer added to a 2010 system. Permissions inherit from the original design. Evidence and audit are secondary. Tool governance is limited by legacy boundaries.
Architecture designed for AI from the start. MCP tool calls. Role-based data access. Evidence citations on every response. Audit row for every agent action.
What AI guardrails exist
Status
The MCP tool call architecture and audit framework are built. Specific investment workflows — portfolio state queries, Decision Packet preparation, regime signal retrieval — are available for selected preview customers. Broader tool coverage and natural language interfaces are on the roadmap.
Next step
Walk through a governed tool call — query, evidence citation, audit row.
Request Demo