JS&C Automation
An interactive demonstration of the PE AI Engineering Platform: retrieval-grounded answers, human review gates and an audit trail, running on illustrative data.
- Next.js
- React
- TypeScript
- Tailwind CSS
Overview
JS&C Automation is the public demonstration of our PE AI Engineering Platform, an AI workspace for financial operations. Instead of describing the pipelines in a deck, the demonstration lets a prospect run representative workflows and watch the review and audit path in real time.
The demonstration is honest about what it is. It runs on illustrative data, has no live client documents and no live external connectors. Every screen carries the label.
What it demonstrates
The workspace explores three patterns we build for clients:
- Reconciliation workflows. Combining statement or position data, identifying variances and producing an exception report for a named reviewer.
- Document review. Ingesting documents, extracting structured fields with cited source passages and routing uncertain extracts to a human reviewer.
- Reporting. Turning approved data into a management or compliance-style report with an audit trail.
Each flow follows the same shape we build into client systems: input, processing, human review, signed output.
The product behind it
The PE AI Engineering Platform is the internal workspace these patterns run on. It is a product in development, demonstrated publicly so that finance teams can judge the interaction model before commissioning a build.
Two design decisions carry through everything:
- Answers cite their sources. Every generated answer links to the passages it came from, so a reviewer can check the work.
- Humans approve. Nothing moves to an output or a report without a named reviewer signing off.
Design & engineering
- Next.js frontend with the platform workspace split across tool surfaces: chat, documents, radar and reporting.
- The JS&C design system: warm cream background, sage accent, serif display type, compact radii on panels, pill shapes reserved for controls.
- The full stack behind the demonstration includes a FastAPI backend with persistent storage, and the frontend proxies API calls through its own routes.
- No emojis, no decorative gradients, provider-neutral copy throughout.
Status
Internal demonstration. The platform continues development in private repositories; the public demo ships with every engagement conversation as a starting point for scoping.
Run the demonstration: automation.jonathansimpson.co.
Outcome: Shipped an interactive demonstration of the PE AI Engineering Platform at automation.jonathansimpson.co. Every screen labels itself as a demonstration running on illustrative data.