Shadow Comp

• 

  Shadow Comp turns a business brief into an evidence-based decision report

• 

  Paste a business brief once; Shadow Comp structures it into editable inputs

• 

  The platform organizes inputs, runs deterministic scenarios, writes analysis, and renders a report.

• 

  AI explains the result while the simulation engine calculates the numbers

Inspiration

Founders often make major pricing, hiring, and growth decisions without a clear way to see their operational impact. Shadow Comp was inspired by the idea of letting teams rehearse their company’s future before making real-world decisions.

## What it does

Shadow Comp turns a business brief into an editable, evidence-based business decision report.

It helps users compare choices such as pricing, usage limits, and hiring plans, then shows their possible impact on revenue, costs, churn, support workload, infrastructure pressure, and runway. It also generates a downloadable PDF report and model-data CSV.

## How we built it

We built Shadow Comp with TypeScript, a web dashboard, and a deterministic simulation engine.

Codex and GPT help organize business briefs, identify missing assumptions, research context, and explain results. Our TypeScript calculation engine handles the financial and operational calculations, keeping generated AI content separate from verified results.

We also built editable report inputs, industry KPI models, scenario comparisons, sourced research, report-confidence checks, PDF generation with PDFKit, and automated tests.

## Challenges we ran into

One of our biggest challenges was PDF generation. We needed to prevent text overflow, blank pages, misplaced footers, and broken layouts while keeping the report readable and professional.

We also ran into incorrect values when assumptions, research, and calculated results were not clearly separated. We solved this by creating clear boundaries between user inputs, AI- suggested assumptions, sourced information, and deterministic calculations.

Some features were also unreliable during development, especially report generation, research availability, and scenario flows. We improved them through validation, fallbacks, timeouts, typed interfaces, and automated testing.

## Accomplishments that we're proud of

We are proud that Shadow Comp does more than generate generic AI advice. It gives users a clear view of what is assumed, what is researched, and what is calculated.

We are also proud of building scenario comparisons, a deterministic simulation engine, downloadable PDF reports, source links, evidence checks, and safeguards against unsupported recommendations.

## What we learned

We learned that AI is most valuable when it structures messy information, identifies missing details, researches context, and explains outcomes—not when it is trusted to invent business results.

We also learned that reliable decision-making tools need strong boundaries between AI output and verified calculations. Accurate values, validation, testing, and clear reporting are essential when helping people make important business decisions.

## What's next for Shadow Comp

Next, we will continue revolutionizing Shadow Comp into a more powerful business decision platform with smarter simulations, richer industry-specific models, stronger research capabilities, and more accurate forecasting.

We plan to support more business types, improve collaboration and report sharing, add richer visual scenario comparisons, strengthen PDF exports, and deliver an even better experience for founders, startups, and decision-makers.

Built With

• chatgpt
• codex
• css
• html
• node.js
• openai
• pdf.js
• railway
• rest-api
• supabase
• typescript
• vitest