The first artificial intelligence agent is an agent for the bill for Qonto customers

when twin Get out of surreptitiously In January 2024, artificial intelligence agents were more theoretical than reality. Today, the Paris -based company will be issued a headquarters for an automation agent in partnership with QontoFintech, which provides business bank accounts for more than 500,000 customers throughout Europe.

If you want to automate frequent tasks, there are already several ways to address these problems. Some companies use automation products based on the interface programming interface, non -code, or low -code, such as Zabir. Others depend on RPA programs, such as UIPATH.

With her team of nine people, Twin believes that there is a more efficient way to deal with automation. It may also have been guessed, as it included artificial and computers.

The invoice operator, the first TWIN product designed for Qonto, is a good example of why artificial intelligence can be used. Qonto pays millions of bills per month. Customers spend several hours a month in collecting and downloading bills on Qonto.

Over the past three months, Twin has created an automatic tool for stabilizing the bill that could accelerate this process. When users operate the invoice player, Twin First brings the list of transactions with lost bills. Then he displays the list of services that need access to downloading bills next to the browser window that displays the agent’s actions.

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

Next, the Twin invoice operator automatically finds the list of previous transactions, download bills, and connects PDFS with transactions in your Qonto account.

“When you do it on the Qonto scale, it mainly needs to cover the very long services. Thousands, tens of thousands, and soon hundreds of thousands of different services that everyone uses,” said co -founder and executive director of Twin Hugo Mercier during a pilot offer of the product.

He added: “This will be completely impossible with RPA because you will need to create a single dedicated text program for each site, and every time the website changes, you will have to modify the text program.”

As for API automation products such as Zapier, Mercier said it took Zapier 10 years to support 8000 requests on its platform. Twin already supports thousands of applications for its bill operator just two months after starting the product.

Behind the scenes, Twin operates a chrome -based web browser on a server. Use the CUA model (Opeenai). In fact, Twain was one of 15 companies that managed to try CUA in Beta.

CUA is also the model that operates the Openai operator, which is the Prosumer product that allows you to enter the promulgation to allow the agent to procedure for you. In addition to better performance, Twin believes it is easier to use agents who browse the web for you.

“We have worked a lot to make the experience very simple. We are really targeting the end user, and perhaps people who are not easy for technology. They do not have to claim or form anything. You just log in to your accounts, run it, and move to find bills,” said MERCIER.

After recovering the bill, Twin believes that there are many industries that can benefit from the B2B agent applications. For example, agents can automatically manage applications for the e -commerce company, classify the market catalog, or recover information for the communication center agents.

Twin puts a future in which artificial intelligence agents become cheaper, faster and more accurate through a wide range of tasks. Now, let’s see whether to start running the basic agent platform that operates the invoice player into a product that developers can start using in their own applications.