Show HN: Visual ‘what if’ financial scenario simulations with GPT-powered AI agents (human in the loop)

We’re trying to create a Financial Jarvis.

Here’s the thesis.

No one in their right mind would trust an LLM to drive a bunch of time-series financial calculations, years out into the future, of any real complexity. Simply too many variables, dependencies, nested calculations.

Plus their job would be on the line if it was incorrect.

Spreadsheets work - sort of - but they lack context. Each cell is just a number - no true metadata. And you can only really generate one scenario at a time.
You have outputs - but limited insights into the when, the what and the why of those outputs. At its core, spreadsheets are not designed with TIME as their primary axis.

To make this work - to provide operational context to a model - you need a different, repeatable data structure for defining time series financial calculations and the business logic. You need to be able to see each calculation and validate the inputs to trust the outputs.

whatifi is a visual, financial and strategic modeling framework that uses GPT and a node-based architecture to help you build (and audit) complex “what-if” scenarios without drowning in spreadsheets.

Think Excel meets decision trees - but in the Multiverse.

Sharing a link to our post today on HackerNews: https://news.ycombinator.com/item?id=42724739