Prototype of an application for window pricing using AI

The goal of the project is to create a prototype (MVP) of an innovative tool for pricing window carpentry (windows, doors, blinds, garage doors) with the auxiliary use of AI.

Currently, there are programs on the market for pricing carpentry that require clicking through many options to get the final price for the product. The pricing process is time-consuming even though 90% of the window specifications are repetitive. Window distributors do not want to use the software offered by window manufacturers because it is too demanding and prefer to send the specifications to the manufacturer for their employee to provide the pricing. The specification sent this way is written in natural language in the email content and contains only the necessary data (product type, quantity, dimensions, color, possible additions), while the rest of the specifications are default (e.g., if glazing is not specified, standard glazing 4/16/4 should be calculated; if color is not specified, white should be calculated, etc.).

I want to propose a convenience where the user (distributor or private individual) enters the product specification in natural language in an online interface. This text is sent via API to OpenAI, which parses the text, extracts the necessary data for pricing (and applies default values in case of missing data) and returns the organized text in JSON format to the program, which connects via API to a spreadsheet containing the price list and calculates the prices. After calculating the prices, the program returns the result to the interface.

I am looking for a budget solution. The aim is to create a basic mechanism to check if customers will use this.