Twin and Qonto: AI Simplifies Corporate Invoice Recovery | Top 10 most popular large language models in the world | Train llm on own data github | A comprehensive overview of large language models | Turtles AI

Parisian startup Twin has developed “Invoice Operator,” an AI agent that automates invoice recovery for Qonto customers, reducing the time spent on this task.

Key points:

• Advanced automation: “Invoice Operator” uses AI to streamline the invoice recovery process.
• Integration with Qonto: The AI ​​agent integrates directly with business bank accounts offered by Qonto.
• Extended support: Twin already supports thousands of applications, expanding the possibilities for automation.
• Strategic collaboration: The partnership between Twin and Qonto highlights the importance of innovation in the fintech sector.

In 2024, Twin, a Paris-based startup, unveiled “Invoice Operator,” an AI agent designed to automate invoice recovery for customers of Qonto, a fintech that offers business bank accounts to over 500,000 businesses in Europe. Traditionally, repetitive tasks have been automated through API-based tools like Zapier or RPA software like UiPath. However, these solutions have limitations, especially when it comes to adapting to a wide range of services and platforms.

With a team of nine, Twin has developed a more efficient solution by leveraging AI. "Invoice Operator" faces a significant challenge: Qonto handles millions of invoices monthly, and customers spend several hours a month collecting and uploading these invoices. The AI ​​agent automates this process by identifying transactions with missing invoices, accessing the relevant services, and downloading the invoices, which are then automatically attached to the transactions in the user’s Qonto account.

Unlike traditional RPA software, which requires custom scripts for each site and continuous updates in case of changes, "Invoice Operator" uses OpenAI’s computer-using agent (CUA) model. This approach allows the agent to navigate the web autonomously, dynamically adapting to changes in site interfaces. Furthermore, while platforms like Zapier took years to integrate thousands of applications, Twin achieved similar support in just a few months.

User experience was a priority for Twin. "Invoice Operator" is designed to be intuitive, even for non-tech-savvy users. After logging into their accounts, the agent proceeds autonomously to retrieve invoices, requiring the user’s intervention only to enter credentials when necessary.

Beyond invoice retrieval, Twin sees broader applications for AI agents in the B2B sector, such as automatic order management for e-commerce, catalog classification in marketplaces, or information retrieval for call center agents.

This collaboration between Twin and Qonto represents a significant step in the integration of AI to improve the operational efficiency of companies.