Why Generic AI Isn’t Enough for Finance Tracking (How Centico AI Connects Binance, Bybit, and Your…

Why Generic AI Isn’t Enough for Finance Tracking (How Centico AI Connects Binance, Bybit, and Your Bank in One Chat)

Financial life isn’t as simple as it seems. Primarily, your personal economy is determined by your habits and behavior. Financially literate people can roughly be divided into two categories: delegators who, having the means, hire experienced financial advisors/agents, and DIY control enthusiasts who manage things their own way, analyze their financial situation, and plan their lives to improve it.

This second category of people used to rely on personal paper records, then started migrating everything to spreadsheets, and later to mobile expense trackers. However, tracking your expenses is just raw information that doesn’t bring much value on its own. In order to analyze your behavior, you still have to spend a lot of time just to make sense of things. Then LMs (Language Models) appeared, allowing you to hand over all your accumulated data for analysis and get near-instant results with insights.

But nowadays, a financially literate person’s financial model doesn’t consist of a single resource and “single-step” operations. It includes dozens of financial resources, dozens of daily transactions, debts, recurring payments, loans, shared budgets, and much more. As a result, managing such a model becomes extremely difficult. It drains a person’s time and mental resources. Even when using standard trackers with a beautiful interface, entering information takes up a lot of time, and extracting actual value from it requires even more effort. Even with modern apps where you can “dictate” an expense, it’s still just tracking, albeit in a new way.

The mission of Centico is to significantly conserve a person’s resources. The vision of this product is to help people multiply their capital using the products of the MEAI ecosystem.

Centico essentially acts as a connector between these cyclical actions. Recording, choosing a category, analyzing, drawing conclusions — the LM does all of this for the user. The sequence shifts like this: a person connects their bank in the app with one click, and all transactions are automatically written to the database. At the moment it is recorded, the AI reacts to these events and, knowing the user’s context, can proactively write a recommendation, alter a plan/strategy, remind you about limits, and so on. It is real-time analysis. We are gradually implementing integrations with various banks both in Ukraine and abroad to reach the widest possible audience and give people the opportunity for full automation.

A knowledge model of the user’s behavior is built from their messages and transactions. This allows for more personalized communication aimed at a realistic user experience, simulating the experience of working with a live financial assistant whom you simply text in a Telegram chat.

[embed]Centico AI: talk about your financial portfolio in a chat

Right from the start, the program asks the user about their financial sentiments — this is the first step toward personalization. Centico’s functionality in its initial versions includes creating resources in a traditional portfolio and connecting cryptocurrency resources (Binance, Bybit, WhiteBit, OKX). In the chat, you can get detailed information on every asset from the connected crypto platforms. All actions available in traditional trackers can be done in seconds right in the chat (logging expenses, income, analysis by days, months, categories, keywords, etc.).

You might ask: You can take bank statements and upload them to ChatGPT, so why use Centico? The answer is quite practical. Here is the number of operations required to analyze expenses that way:

1. Open your banking app and download an account statement (usually just for one account).
2. Go to the chat and upload this statement.
3. Describe the context, your goal, and the conditions.
4. Get an answer regarding this specific resource and move on.

And you might have 5 or 10 such resources… And if we’re talking about cash, you definitely can’t generate a statement for that. Furthermore, you will NOT get maximum personalization, because that requires: history, knowledge about you and your habits, and full context (which you definitely won’t explain to GPT in every new prompt when you want to discuss something specific, like a financial strategy. For it to work correctly, you’d have to update the information from the statements every single time, as well as add new transactions that occurred outside of them).

In Centico AI, this is available right out of the box. It can be accessed at any second without taking a huge number of steps to get the necessary information; you can analyze by any category or even by keywords from transaction details. Most importantly, it is always based on the user’s up-to-date transaction state and personal context.

What are the downsides of using Centico?

• Integrations with banks take time; not all banks provide an open API (an interface for connecting third-party systems), which makes cooperation with them impossible.
• LMs can hallucinate, which is rare but can potentially lead to errors.
• Over time through communication, context changes, and sometimes you need to remind the AI about it (for example, that the price of an upcoming planned purchase isn’t the full 11,000, but 7,000).
• You need to formulate your messages correctly and not mislead the model, because the same rule applies to it as to a human: it will understand exactly how you explain it.
• In the current implementation, Centico AI only messages you first when a specific event occurs.

These drawbacks, while not always pleasant, are very similar to human interaction, and it is within the user’s power to communicate with the AI in a way that allows it to understand them correctly and provide valuable insights.

Conclusion: Thanks to the availability of transactions and context gathered over half a year, the LM forms a clear picture of the user: what they like, where the budget gaps are, what to avoid, and what habits to stick to. The user can make deep, multi-year plans for purchases and vacations, and understand when it’s necessary to push themselves for higher earnings or push management at work for a raise because their current situation doesn’t meet their financial expectations. When you don’t have to spend time manually calculating “how much money will I have in 7 months and 13 days” your stress levels decrease. Absolute clarity in financial matters makes a person happier!

References

Download links.

Centico AI for iOS

Centico AI for Android