Working with Bosbec and AI

This series of articles guides you through a structured approach to building a simple web tool using Bosbec and AI. In this example, we will create a basic to-do list application, but the principles would be useful regardless of what your project might be.

The backend will be implemented in Bosbec using HTTP-in triggers, and the user interface will be generated with the help of AI. In the series, we’ll use Visual Studio Code and GitHub Copilot to generate code, but you can use any agent you prefer. While using an IDE and GitHub has its benefits, simply using the ChatGPT prompt window is enough for the scope of this project.

Before you begin, it is recommended that you document your thoughts about the project. Open a text document and write down your answers to the questions below, along with any initial considerations. This will help you clarify the scope and requirements before implementation.

1. Define the Problem

Start by clearly identifying the problem you want to solve.

In this case, the objective is to build a simple interface where users can:

• Create a to-do list
• Add tasks to the list
• Mark tasks as completed

Defining the problem explicitly ensures that development remains focused and avoids unnecessary complexity.

2. Identify the Required Components

Next, determine which components the system requires.

A practical method is to visualize the finished tool and describe what you expect to see. For example:

• A to-do list
• A set of tasks
• A way to track completion status

Based on this, you can design the data structure before building any workflows. Below is an example JSON structure representing a to-do list:

{
    "id": "b7b4bf27-9c71-4204-ac99-887e452a37ea",
    "title": "Today's tasks",
    "createdOn": "2025-02-27 15:00:00Z",
    "tasks": [
        {
            "id": "a95b6a3f-7da7-4c91-9718-e66fbe4eab3e",
            "label": "Buy groceries",
            "completed": "false",
            "createdOn": "2026-03-05 13:23:51Z"
        },
        {
            "id": "7004a586-9d01-41f1-86ca-5bbf1530f866",
            "label": "Finish project report",
            "completed": "false",
            "createdOn": "2026-03-05 13:24:02Z"
        }
    ]
}

Structure Explanation

The list object contains:

• id – A unique identifier used to reference specific lists
• title – The name of the list
• createdOn – A timestamp used for sorting or filtering
• tasks – An array containing task objects

Each task object contains:

• id – A unique identifier
• order – Used for sorting tasks within the list
• label – The task name
• completed – A boolean indicating completion status

Using AI to Generate Structure

You can use your AI assistant to generate an initial data model. For example:

“Can you provide a JSON structure for a to-do list? I need a list name, creation date, and an array of tasks.”

You can then refine the generated JSON or use it as-is during early development, adjusting it as new requirements emerge.

3. Define Required Functions

Once the data structure is defined, identify the functional requirements. For this example, we’ve decided that the following operations are needed:

To-Do List

• Create list
• Delete list
• Rename list

Tasks

• Add task
• Delete task
• Rename task
• Reorder task
• Mark as completed
• Mark as not completed

Document these functions and use them as a development checklist when building your workflows in Bosbec.

Summary

By:

1. Clearly defining the problem
2. Designing the data structure
3. Listing required functionality

…you establish a structured foundation before implementing workflows. This preparation reduces rework, clarifies scope, and enables a more systematic build process within the platform.

Next up, we’ll use the document we’ve written to start building our backend API.