DEV Community

Cover image for Building a Multi Platform AI Budget Coach
Kevin Coto🚀💡
Kevin Coto🚀💡

Posted on • Edited on

Building a Multi Platform AI Budget Coach

The Problem Was Me

I was making more money but my savings were not going up. Every raise just meant I spent more. I tried every budgeting app on the market. They all worked for a week then I forgot about them.

The pattern was always the same. Download app, enter expenses for three days, get bored, uninstall. The apps required too much work. I had to open them, navigate menus, fill forms. When I was tired after work, I was not going to do that.

I realized the problem was not my spending. It was the tools. If budgeting required effort, I was not going to do it consistently.

Meet People Where They Are

I wanted to build something that required zero effort. No app to open, no dashboard to check, no forms to fill. Just send a message like you already do hundreds of times a day.

But people use different apps. In some countries everyone is on WhatsApp. In others it is Telegram. Developers live on Discord. If I picked one platform, I would miss everyone else.

So I built for all three. Same backend, same AI, same budget logic. The only difference is the messaging adapter on top.

The Moment It Clicked

The first time someone typed "spent 45 on dinner" and the bot replied with their remaining budget, their reaction was genuine surprise. It felt like magic because there was nothing to learn. No onboarding, no tutorial, no setup. Just type what you spent and it works.

That is when I knew the concept was right. Not because the technology was impressive, but because it disappeared. The user did not feel like they were using a budgeting tool. They were just texting a friend who happened to track their money.

How It Works

FastAPI handles the backend. Each messaging platform has its own adapter that normalizes messages into a common format. The NLP layer parses natural language and extracts amounts, categories, and intent. PostgreSQL stores everything.

The user types "spent 45 on groceries" and the bot logs it, updates the category budget, and replies with a confirmation and remaining balance. No forms, no fields, no friction.

Where It Is Now

DIA works across Telegram, WhatsApp, and Discord. It handles expense logging, budget tracking, savings goals, and spending analytics delivered as weekly digests. The natural language understanding improves over time as it learns how each user talks about their money.

Top comments (0)