Bison Bank AI Bank Assistant

The description states that Bison Bank AI Bank Assistant is an AI chatbot assistant built for banking and financial questions. It claims to use a text-only reasoning model (gpt-oss-120b) in a split-stack architecture with Supabase for data storage and retrieval.

It builds on a Grounded Caption Retrieval-Augmented Generation (RAG) pipeline, where:

• The frontend UI is built using ChatGPT, React, and Bolt.
• Financial documents and charts are stored via Supabase Storage.
• Mathematical logic is extracted from text context and fed into the AI model for processing.

The system supports calculations like WACC (Weighted Average Cost of Capital), using structural data packets to compute precise answers without relying on visual scanning.

Inference The product is described as a prototype or proof-of-concept, not a commercial offering. It is not evidenced that it has been deployed beyond the hackathon submission.

The author positions the assistant as an AI knowledge hub for institutional banking data, aiming to reduce time spent manually analyzing financial sheets and charts.

Key claims:

• The system answers high banking math questions.
• It dynamically searches through previously logged financial data.
• It calculates complex financial models like WACC using text-based reasoning.
• It avoids multimodal vision models by leveraging structured database tagging.

Inference The positioning is focused on institutional use cases, but there is no evidence of a target market or customer base beyond the authors’ own description. The claims are aspirational and not validated.

The project description states that the assistant is built for "institutional banking data" and aims to help users analyze balance sheets and financial charts.

Inference The target customer appears to be professionals or institutions working with financial data, but no specific customer segments or personas are defined. No evidence of early adopters or pilot customers is provided.

There is no evidence in the description of a business model or pricing structure. The project is described as a hackathon submission and lacks any indication of monetization, licensing, or sales channels.

Inference No commercial business model is evident from the self-reported description.

The system uses:

• Frontend: React, ChatGPT, Bolt, Vite
• Backend: Supabase (Storage + relational DB)
• AI Model: NVIDIA NIM gpt-oss-120b (text-only)
• Architecture: Split-stack with RAG pipeline
• Data Handling: Client-side PDF parsing, database keyword matching (.ilike), and structural data feeding into the model

Inference The technical stack is described as a prototype or hackathon implementation. No evidence of scalability, production deployment, or robust infrastructure is provided.

The project was submitted to the OpenAI 2026 hackathon. It has no evidence of:

• Revenue
• Customers
• Product usage metrics
• Live deployment
• Market traction

Inference No traction or maturity signals are evident beyond the self-reported hackathon submission.

The description does not mention any competitors or market positioning relative to existing AI banking tools or financial data platforms. It is unclear whether similar solutions already exist in the market.

Inference No competitive landscape is described, and no evidence of prior market analysis or differentiation is provided.

• Unverified claims: All features and capabilities are self-reported without independent validation.
• No traction or revenue: The project is only a hackathon submission with no evidence of adoption or monetization.
• Prototype nature: No indication that the system has moved beyond prototyping or testing.
• Limited team size: Only two members, which may limit execution capacity.
• Unclear commercial viability: No business model or go-to-market strategy is evident.

1. What specific financial use cases are you targeting in the institutional banking space?
2. Have you tested this system with real financial data or users beyond the hackathon?
3. Is there any plan to move beyond a prototype into a product or service?
4. How do you intend to monetize this solution, if at all?
5. What is your roadmap for scaling the system beyond the current architecture?