- Joined
- Jan 20, 2026
- Messages
- 345
- Reaction score
- 2,751
Attempts to prove to the system that the loan is already closed turns into a real test of patience.
And is increasingly taking on the work that people used to do, and now it is the turn to the collectors. American companies began to massively connect voice bots to call debtors, and automatic systems are already making millions of calls a month. Together with the increase in the number of delays, the number of complaints about errors of such services is also growing.
About the distribution of AI-agents among collection companies told edition of Wired. The reason for the material was the story of a resident of Seattle named Ben, who was called by a virtual Eve assistant from ProCollect. The bot demanded to pay a debt of $ 266 in the old dispute with the landlord, although the issue has long been closed.
During the conversation, the voice agent immediately offered to pay the debt with a card or bank transfer and for a long time did not switch the call to a live employee. Ben decided to check how far the algorithm is able to go, and began to ask non-standard questions about the structure of payments. Instead of transferring to the operator, the system continued the dialogue.
According to Ben, the conversation gradually turned into an absurd role-playing dialogue, after which the call was still transmitted to a person. The employee confirmed that the debt had already been repaid and there were no complaints against the interlocutor.
The co-founder of the company Altur Pedro Fernandez told Wired that collection agencies have become one of the most active clients of developers of automated call centers. The Altur system, he said, makes more than 2.5 million calls to debtors every month.
The problem is the quality of the data used by collectors. Information about debts often shifts from one lender to another, stored in huge tables and regularly contains errors. With the participation of live operators, the dispute situations can still be quickly clarified, but automatic agents often continue to demand payment even after the debt is repaid.
Against the backdrop of a high level of private debt in the United States, the market for such services continues to expand rapidly. At the same time, the automation of collection calls already raises questions not only to the quality of AI, but also to how easy it is for a person to achieve bug fixation.
And is increasingly taking on the work that people used to do, and now it is the turn to the collectors. American companies began to massively connect voice bots to call debtors, and automatic systems are already making millions of calls a month. Together with the increase in the number of delays, the number of complaints about errors of such services is also growing.
About the distribution of AI-agents among collection companies told edition of Wired. The reason for the material was the story of a resident of Seattle named Ben, who was called by a virtual Eve assistant from ProCollect. The bot demanded to pay a debt of $ 266 in the old dispute with the landlord, although the issue has long been closed.
During the conversation, the voice agent immediately offered to pay the debt with a card or bank transfer and for a long time did not switch the call to a live employee. Ben decided to check how far the algorithm is able to go, and began to ask non-standard questions about the structure of payments. Instead of transferring to the operator, the system continued the dialogue.
According to Ben, the conversation gradually turned into an absurd role-playing dialogue, after which the call was still transmitted to a person. The employee confirmed that the debt had already been repaid and there were no complaints against the interlocutor.
The co-founder of the company Altur Pedro Fernandez told Wired that collection agencies have become one of the most active clients of developers of automated call centers. The Altur system, he said, makes more than 2.5 million calls to debtors every month.
The problem is the quality of the data used by collectors. Information about debts often shifts from one lender to another, stored in huge tables and regularly contains errors. With the participation of live operators, the dispute situations can still be quickly clarified, but automatic agents often continue to demand payment even after the debt is repaid.
Against the backdrop of a high level of private debt in the United States, the market for such services continues to expand rapidly. At the same time, the automation of collection calls already raises questions not only to the quality of AI, but also to how easy it is for a person to achieve bug fixation.
