An investment firm ran its portfolio on broker statements, exports, and spreadsheets. Positions arrived in different formats from different brokers, options needed their own handling, and every risk question waited for someone to rebuild the picture by hand. The firm wanted software built around how it invests, with its data in a shape AI tools could use.
We built the software in stages. Parsers normalize broker and options files into one position model, with each ticker mapped to its data source. A portfolio application handles the day-to-day changes, and custom risk dashboards show exposure the way the firm's principals review it. Power BI covers the reporting the team wanted to build for itself. The latest addition is an MCP server, built on the Model Context Protocol, that connects the firm's AI assistant to the firm's own API, so a principal can ask a question in plain language and get an answer drawn from the firm's data through the same API the applications use.
A small team carried the work for more than a year: a lead engineer and two developers shipping production releases as the firm's needs came up. Working in small releases let the firm add capabilities, from new risk views to the AI connection, as continuing work under one engagement.
The firm sees its whole portfolio in one model instead of across broker statements, and risk questions are answered from dashboards instead of rebuilt spreadsheets. The MCP server extends that same data to the firm's AI assistant. That last step depends on everything before it, because an AI assistant can only answer well from data that is already clean, modeled, and behind an API.
Client details in this case study are generalized, and in places combined across engagements, to protect confidentiality. The work draws on our fintech software development and AI integration practices. Our article on MCP security covers how to connect AI tools to internal systems safely, and our analysis of custom software versus off-the-shelf covers when building wins. We are glad to walk through comparable work under NDA.
