Architecture · WattIQ EPM Decision Agent
A governed AI that reads the World Bank–ESMAP Electricity Planning Model (EPM), reasons under a trust-tiered policy, cites every figure to its source, and improves from use. These pages explain the whole system — pick a view.
A knife-style, self-contained slide deck — the raw .GDX → charts, graphs & insight story, with live proof-of-concept charts drawn from real Bangladesh figures, country-context grounding, and a reconciliation check on every number. It now mirrors the product's stronger visual language: spotlight overlays, scene-by-scene storytelling, and a decision landing. Light/dark, keyboard-driven, projector-ready.
The reasoning explained — the 7-stage pipeline, the 12-step reasoning workflow inside it, how it grounds every figure in the ESMAP EPM GAMS source, and the trace-before-trust rule that governs it all. The best page to email or brief with.
OpenWhat you can do on the platform, one page — ask, set your role, build a Charting Station gallery, present in Theater, export branded reports, and download scene-based decks. Print-friendly.
OpenAuto-plays the pipeline — or click any stage to reveal its tech stack. It now reflects the expanded presentation layer: Charting Station, gallery storytelling, and scene-aware exports.
OpenThe full blueprint — inputs, grounded EPM analysis + provenance, the LLM reasoning loop, display & export, the learning loop, and the platform. Every system and its role.
OpenThe value narrative on one landscape page — grounded → reasoned → cited → validated — now extended with Charting Station, gallery storytelling, and scene-preserving exports. Print-ready for a board pack.
OpenEvery layer and library — exact Python packages, JavaScript libraries, platform services, data formats, chart forms, and the model.
Open