Twin’s first AI agent is an invoice agent for QONTO customers

When Twin It came out of Stealth in January 2024, AI agents were more theoretical concept than a reality. Today, the Paris -based company releases an automation agent in collaboration with QontoThe start of Fintech offering business bank accounts to more than 500,000 customers across Europe.

If you want to automate recurring tasks, there are already several ways to deal with these problems. Some companies use automation products based on API, non -code or low code such as Lively. Others rely on RPA software, such as Uipath.

With the group of nine people, Twin believes that there is a much more effective way to handle automation. As you may have guessed, it included artificial intelligence and computer use models.

The invoice operator, the first Twin product designed for Qonto, is a good example of it because it is reasonable to use artificial intelligence. Qonto handles millions of invoices per month. And customers spend several hours a month gathering invoices and uploading them to Qonto.

During the last three months, Twin has created an automatic invoice tool that can accelerate this process. When users start the invoice operator, Twin First transfers the list of missing invoices. It then shows the list of services that needs to be accessed to download the invoices next to a browser window that shows the agent’s actions.

If you need to log in to a service to download the invoices, the browser ceases and asks you to enter your credentials manually. Once you do this, you can click on a button to let the agent continue his work.

After that, the TWIN invoice operator automatically finds the list of past transactions, invoice downloads and connects PDFs to Qonto account transactions.

“When you do this on the Qonto scale, you basically need to cover a very high tail of services. Thousands, tens of thousands and soon hundreds of thousands of different services used by everyone,” said Gemini Co -founder and CEO Hugo Mercier during a product.

“And this would be completely impossible with RPA because you will need to create a single custom scenario per website and then each time the site changes, you will have to modify the scenario,” he added.

As for API -based automation products such as Zapier, Mercier said it took 10 years to Zapier to support 8,000 applications on its platform. Twin already supports thousands of applications for the invoice operator just two months after the product is launched.

Behind the scenes, Twin runs a Chromium -based web browser on a server. The start uses Openai’s OpenAi (Agent-Aging Agent) model. In fact, Twin was one of the 15 companies that had to try the Cua on beta.

CUA is also the model authorizing the Openai operator, the Prosmer’s product that allows you to enter a question to let an agent perform an action for you. In addition to better performance, Twin believes it should be easier to use agents that browse the web for you.

“We worked a lot to make the experience extremely simple. We are really aiming for the end user. Maybe people who are not technology -friendly, do not need to encourage or shape anything, just log in to your accounts, start it and navigate to find the invoices,” Mercier said.

After recovering an invoice, Twin believes that there are many industries that could benefit from B2B Agentic applications. For example, agents could automatically manage orders for an e -commerce company, classify the list of a market, or retrieve information on call center agents.

Twin places a future in which AI agents become cheaper, faster and more accurate in a wide variety of duties. Now let’s see if the start can turn the key agent who authorizes the invoice operator into a product that developers can start using in their own applications.