Ask for the answer.
The engines do the running.
The applications disappear behind a conversation. You type what you want in plain language; an agent plans the work, drives Imagio, Conversio, Præsagio, Punctio, Maturatio and Estimatio through the engine API, and comes back with results — every number computed by an engine run, never invented by a language model.
No hallucinated economics
The language model never produces a number. Every volume, rate and NPV on screen is the output of the same deterministic engines the desktop apps run — the agent only decides which one to call.
The chips are the receipt
Each reply shows which engines ran and with what scope; result packages attach to the chat and re-run identically outside it. If a number matters, you can trace it to its run.
Agents sweep, people decide
Agents are tireless at running cases, not entitled to conclusions. Concepts, rankings and deltas queue for your geoscientists' judgement — the same discipline as the Malaysia bid round.
The same engine API, wherever you deploy it.
The chatbot is a front door, not a separate product. It drives the identical engine API that ships with an enterprise deployment — inside your perimeter, against your database, under your identity — and it inherits every customization: your fiscal presets, your risking matrix, your reservoir physics. Point it at one surface for a session like the one above, or at the whole archive overnight.
