A founder setting up a new architecture firm from the ground up didn't want to bolt operations onto off-the-shelf tools as an afterthought — he wanted project management, time tracking, and client communication built AI-native from day one, specifically so the system could do more than track work: analyze it. The goal was a system that could generate quotations and proposals automatically, tell him where time was actually going on a given project, and give clients live visibility into their own projects without manual updates.

Full PM platform (Angular frontend, .NET backend) — complete task management, time-sheet tracking, and client-facing dashboards, architected and built end-to-end. Clients get their own dashboard view into their project's status rather than waiting on manual updates from the firm.

AI service layer (Python, Azure AI) — a dedicated Python service connected to Azure AI, wired into the platform's data (tasks, time entries, project history) so the AI layer would have what it needed to reason about a firm's real work. This is the layer automatic quotations, proposal generation, and smart time-sheet analysis were designed to run on.

Where the build stood at close: the full platform was completed and functional. The AI-native feature set itself — the quotation generation, proposal generation, and time-sheet analysis logic — was the next phase, with the service layer and infrastructure already in place to support it.

Most firms bolt AI features onto an existing tool after the fact. This system was designed the other way: the operational data model came first, specifically so the AI layer would have real, structured data to reason over instead of guessing from unstructured inputs — the harder, more valuable design decision, and why the AI infrastructure being in place is real, usable groundwork rather than a stub.