Development of a Web platform for fleet management for car rental
Goals - Replace olx for lead generation, paper accounting, create a unified system for searching, signing contracts, renting, payments, debt accounting, with scalability options.
What needs to be implemented step by step.
A) Login via Email
B) User roles
C) Access rights restrictions
Roles: Owner, administrator, service manager, driver, accountant (Built-in accounting system)
After logging in, the user should see: Number of available cars, cars rented, in repair, overdue payments, maintenance, driver debts, recent events, tasks, calendar, income and expense schedule (for each car)
Car card: VIN, license plate, make, model, year, photo, mileage, status, next maintenance, insurance and until what date. Tabs: Repair history, accidents, income, expenses, profit, documents, GPS (Prepare architecture)
Card: Full name, phone, Viber (TG) Photo, driver's license, passport, balance, debt, deposit, rating, blacklist.
History: Rentals, payments, fines, accidents, documents, comments
Creating a rental, fields: car, driver, start date, end date, extension, rental cost, deposit, comment
After creating a rental, the car receives the status “Rented”, a debt is created, and the record goes into history.
Functions: accepting payments, cash, non-cash, reducing debt, transaction history, cash register.
Creating a repair, fields: car, description, mechanic, labor cost, parts cost, receipt, date, status, after closing the repair, the car's expenses are automatically updated.
The section should contain: income, expenses, profit, cash register, debts, charts, for each car.
Any change must be recorded. Fix: who changed, when, what was changed, old value, new value.
Prepare architecture for: push, Telegram, Email
- Telegram AI Assistant (text)
This is a key functionality of the project.
It is necessary to implement a Telegram AI Assistant that can understand user text messages, with the possibility of implementing the assistant in the CrM system itself, as an operator assistant.
Examples: “Book Aveo 3661 for Sergey from Monday”, “Sergey paid and extended the rental 4500 for a week”, “Set the value of Lacetti 7710 in repair” analysis questions “Who needs to pay today?” “Which cars are available and ready?”, “Which car is the most profitable/unprofitable?”
Logic of operation:
Receive a message, pass it to OpenAI, determine intent, find entity (Object, driver, car, amounts, etc.), if necessary, ask a clarifying question, call the corresponding API function, send the result to the user. It is important to use Function Calling/Structured Outputs, not text parsing with regular expressions.