Twin’s first AI agent is an invoice-retrieval agent for Qonto customers
When Twin Came from Stealth in January 2024, AI agents were more a theoretical concept than a reality. Nowadays, the Paris -based company releases an automation agent in collaboration with QontoThe FinTech Startup that offers business bank accounts to more than 500,000 customers throughout Europe.
If you want to automate repetitive tasks, there are already different ways to tackle these problems. Some companies use API-based, no-code or low-code automation products such as Zapier. Others rely on RPA software, such as Uipath.
With his team of nine people, Twin thinks that there is a much more efficient way to handle automation. As you may have guessed, it was because of models for artificial intelligence and computer use.
Invoice operator, the first Twin product designed for Qonto, is a good example of why it is useful to use artificial intelligence. Qonto handles millions of invoices per month. And customers spend a few hours a month collecting invoices and uploading to Qonto.
In the past three months, Twin has created an automatic invoice tool that can speed up this process. When users start the invoice operator, Twin first picks up the list of transactions with missing invoices. Subsequently, the list of services it needs to gain access to downloading invoices in addition to a browser window with the agent’s promotions.
If you have to log in to a service to download invoices, the browser pauses and you ask to manually enter your references. Once you do this, you can click a button to let the agent continue.
After that, the Twin invoice operator will automatically find your list of previous transactions, downloads invoices and adds the PDFs to the transactions in your Qonto account.
“If you do that on the scale of Qonto, you should basically cover a very, very long tail of services. Thousands, tens of thousands and soon hundreds of thousands of different services that everyone uses,” said Twin co-founder and CEO Hugo Mercier during a demo of the product.
“And that would be completely impossible with RPA because you should make a single adjusted script per website and every time the website changes, you should adjust the script,” he added.
Regarding API-based automation products such as Zapier, Mercier said that Zapier took 10 years to support 8,000 applications on its platform. Twin has already supported thousands of applications for his billing operator only a few months after starting the product to the product.
Behind the scenes, Twin runs a chrome -based web browser on a server. The startup uses the CUA model of OpenAI’s CUA (computer use). Twin was even one of the 15 companies that CUA could try in Beta.
CUA is also the model that the Operator of OpenAi, his Prosumer product, Power Poemert with which you can enter a prompt to have an agent perform an action for you. In addition to better performance, Twin believes that it should be easier to use agents who browse the internet for you.
“We have worked a lot to make the experience extremely easy. We really focus on the end user, perhaps people who are not technically friendly. They don’t have to call or configure anything. You simply log in to your accounts, start it and navigates to find the invoices,” Mercier said.
After collecting invoices, Twin thinks there are many industries that can benefit from B2B -agent applications. For example, agents can automatically manage orders for an e-commerce company, classify the catalog of a marketplace or collect information for call center agents.
Twin throws a future in which AI agents become cheaper, faster and more accurate over a wide range of tasks. Now let’s see if the startup can change the core agent platform that the invoicer of the invoicing in a product that developers can use in their own applications.
