02 · Selected work
ConnectX
Conversation determines intent. The dashboard remains the workspace: an AI-native product for energy across commercial buildings.
- Year
- 2025
- Type
- 0→1 · Building energy
- Role
- Product definition & demo
- Focus
- Intent-led workspace
The tension
ConnectX arrived with personas, KPIs, and AI-generated concepts of a dashboard and digital twin. The ambition was clear. The product model was not. Facility, finance, and sustainability teams needed different answers from the same energy system, and the interface could not permanently show every possible view.
Two directions were credible: a fixed dashboard with a chat assistant, or letting the request invent the screen. Neither was enough. A fixed dashboard cannot adapt across roles. A chat reply cannot hold persistent context or give direct control.
What I designed
The product did not replace the dashboard with chat. Conversation established what someone wanted to understand. The product then assembled a persistent workspace around that intent (KPIs, assets, charts, and twin) that they could keep using by hand.
- Eight bounded scenarios (status, forecast, asset performance, operating conditions) so the model could adapt the workspace without inventing it freely.
- A flow of intent → reasoning → classification → tools/APIs → modules, not an open-ended chat overlay.
- An abstract 2.5D twin that explains the energy system (which assets, what state, which one matters now) instead of counting windows.
Where it stands
This was product definition and a working demonstration, design and engineering together, because the interaction model depended on what could actually be rendered and called. It is not a claim of live telemetry, adoption, or operational savings.