📊 Stratify: Your AI Consulting Copilot

💡 Inspiration

The barrier to entry for top-tier management consulting frameworks (like those used by the Big 4 or MBB) is incredibly high. Startups, hackathon teams, and small businesses often lack the strategic rigor required to validate their ideas, model their finances, or diagnose complex operational issues. We wanted to democratize elite consulting by building an AI copilot that goes beyond conversational chat to actively build structured, professional-grade deliverables.

⚙️ What it does

Stratify is a Generative Strategy Engine designed to be your virtual "Big 4" consulting partner. It adheres to CBS case competition standards to help users break down problems and build strategies through several specialized modes:

Strategy Mode: Automatically generates slide decks featuring professional "Action Titles" and supporting "Kickers."
Diagnosis Mode: Deconstructs complex business problems into MECE (Mutually Exclusive, Collectively Exhaustive) recursive Issue Trees.
Diligence Mode: Provides a comprehensive Risk Dashboard with interactive sliders to stress-test your business assumptions.
Financial Engine: Features a real-time "Valuation Sandbox" for dynamic financial modeling.
Export: Seamlessly packages the generated strategy into PPT format pitch decks for immediate use.

🧪 The Math: Dynamic Valuation Sandbox

To ensure the Financial Engine provides rigorous modeling rather than just estimates, we built a dynamic Discounted Cash Flow (DCF) calculator directly into the sandbox.

The engine calculates the Present Value ($V_0$) of a business strategy using the following formula:

$$V_0 = \sum_{t=1}^{n} \frac{CF_t}{(1 + r)^t} + \frac{TV}{(1 + r)^n}$$

Where:

$CF_t$: The projected Cash Flow for year $t$, generated by the AI based on your business model inputs.
$TV$: The Terminal Value of the business at the end of the projection period $n$.
$r$: The risk-adjusted discount rate. This variable is dynamically influenced in real-time by the user's inputs on the Diligence Mode risk sliders.

🛠️ How we built it

Frontend & UI: We built the application using Next.js 14 and TypeScript for a robust, server-rendered experience. The interface utilizes a custom "Obsidian Executive" dark-mode design system, styled with Tailwind CSS and brought to life with fluid Framer Motion animations.
Data Visualization: We integrated Recharts to visualize the financial models and risk metrics dynamically.
Generative Architecture: We implemented a "Schema-First" Generative Flow utilizing the Tambo SDK. This was crucial for forcing the AI to output highly structured data rather than raw text.
Validation: We used Zod to strictly type and validate the AI's JSON outputs before rendering them into complex UI components like the recursive Issue Trees.

🚧 Challenges we faced

Structured Outputs: Forcing an LLM to generate a perfect, multi-level recursive Issue Tree that directly maps to a React UI component was incredibly difficult. We had to rely heavily on schema-first prompting and Zod validation to prevent the UI from breaking.
Persona Constraints: Ensuring the AI maintained a strict, professional "Big 4" consulting persona without hallucinating irrelevant startup advice required extensive prompt engineering and tuning.

🧠 What we learned

We learned how to effectively constrain generative AI to produce deterministic, structured UI components. We also deepened our understanding of complex financial modeling and discovered how to represent intricate risk metrics interactively on the web using Next.js and Recharts.

🚀 What's next for Stratify

Enhanced Exports: Improving the PPT export functionality to support native, editable charts directly within PowerPoint.
Live Collaboration: Adding multiplayer capabilities so multiple founders can work in the "Valuation Sandbox" and adjust the Diligence sliders simultaneously.
Data Integration: Connecting to live market APIs to automatically pull industry-standard discount rates ($r$) and comparable company multiples.