AI Software for Overdue Debt Recovery

Overdue debt recovery is the hardest work a contact centre does. Volume is high, the conversations carry an emotional load, and every call has to be run with the same discipline. TAHSİLDAR is an AI software built to address all three at once.

Why is the overdue process hard for a human team?

Volume, emotional load and consistency

Volume. As the delinquent portfolio grows, the number of people to call rises quickly. Growing the team, by contrast, depends on hiring, training and time to competence; it does not keep pace with the portfolio.

Emotional load. A collections call is uncomfortable for the person receiving it. Agents face objections, defensiveness and at times a sharp reaction all day long. That load leads to burnout and higher turnover, and every agent who leaves takes accumulated experience with them.

Consistency. Identity verification, the legal notice and the balance disclosure have to be complete in every conversation. In a human team that is a standard maintained only through continuous monitoring and training.

How was TAHSİLDAR designed for this process?

Identity first, then the debt

TAHSİLDAR confirms by voice that it is speaking to the right person before sharing any detail of the debt. The order is not incidental: disclosing debt information to the wrong person is both a privacy breach and a reputational risk. The verification logic is built from the institution's security and compliance rules; it is not a fixed template.

A clear statement of the balance

Once verification is complete the debt is explained plainly: amount, due date, delinquency status, any accruing interest and the available payment channels. Vague wording is avoided; the aim is that the customer leaves the call knowing how much, by when and through which channel they will pay.

Not a script, but flexibility bound by rules

The difficulty of a collections conversation is that everyone's situation differs. One person is paid in two weeks, another can make a partial payment, another disputes the balance outright.

TAHSİLDAR does not read from a fixed script. It understands what the customer says and produces a response within the boundary the institution allows. That boundary matters: the flexibility is the institution's defined rule, not the agent's discretion. Which due date ranges are acceptable, which instalment structures are possible, and in which situations the call is handed to a person are all decided in advance.

How is a payment plan set up?

A process that concludes during the call

When the customer makes a promise to pay, that promise is recorded. If the institution's rules allow it, the plan is created during the conversation and written to the collections system immediately. A written confirmation can also be sent to the customer; we cover running that flow across channels on the omnichannel collections page.

Measuring how well the process works is a separate discipline; our success metrics page sets out which indicators to watch.

For how TAHSİLDAR works end to end, see the TAHSİLDAR page.