VantedgeAI Docs

Core Concepts

The mental model behind VantedgeAI — workspace, modules, agents, approvals, and MCP connections.

A few concepts recur across every module. Learn them once and the rest of the platform reads consistently.

Workspace

Your workspace is your firm's home. It holds the cross-module surfaces you use daily — Home, Pipeline, Portfolio, Tasks, Updates — and the modules your firm has enabled. Navigation is consistent: a left rail groups everything by area, and the active module expands to show its pages.

Modules

A module is a capability area mapped to part of the fund lifecycle — Deal CRM, LP CRM, Data Rooms (ODD/VDR), Risk, CFO Center, Transactions, Pricer, CIM Builder, and more. Enabling a module adds its surfaces to your workspace and makes its agents available.

Agents

Modules ship pre-built agents that automate their busywork — screening a pitch, running diligence, scanning for covenant breaches, drafting LP updates. You start an agent from a module (often a single action like "Run diligence"), and it produces an agent run.

Agent runs & approvals

Every agent execution is an agent run you can inspect — its steps, outputs, and the actions it proposes. Actions that change state route through Approvals, so a human signs off before anything is committed. Agent Runs and Approvals live under AI Operations and are shared across modules.

MCP connections

VantedgeAI is MCP-native (Model Context Protocol). You connect external systems — Bloomberg, Goldman Sachs, Morgan Stanley, Salesforce, QuickBooks, SharePoint, Egnyte, Google Drive, email — as MCP connections, and agents read from them directly. This is also why these docs are agent-readable: the same content humans see is exposed to agents as markdown (see For agents).

Audiences & access

Most surfaces are for your firm's team. LPs get a scoped Investor Portal with only what's shared with them. Developers integrate external systems through MCP connections.

On this page