AI in Credit Management – From Implementation to Measurable Results

Insights from NPL Global 2026 on how data, technology, and professional expertise can strengthen credit management.

AI is becoming part of everyday business operations. The bigger challenge is documenting the improvements the technology brings: better decisions, more efficient debt collection, lower costs, or a better customer experience.

That question was central to the panel discussion “NPL Technology & Innovation” at NPL Global 2026, which was held in London on September 29–30. Jonny Lingvall, Managing Director of Collectia in Sweden and Finland, moderated a discussion with participants from Synthflow AI, BROCC, Revolut, and Accumin Intelligence.

Collectia also shared findings from our CFO Outlook 2026. The survey and discussions at the conference highlighted a specific challenge: How can investments in AI be translated into results that companies can measure and that benefit customers?

Three Key Takeaways from the Discussion

• Reliable data and an understanding of the business provide AI with the necessary context.
• Success criteria must be established before the pilot project begins, and the solution must be tested in day-to-day operations.
• Technology and human judgment must work together, especially in complex customer situations.

Widespread use of AI and limited evidence of its value

In the Collectia Group CFO Outlook 2026, we surveyed 150 senior financial decision-makers in Denmark, Norway, Sweden, and Germany. Among the respondents, 94% use AI in their finance departments, and half report that the technology is integrated into core processes. Nevertheless, 61% are unable to clearly demonstrate the value of AI.

The figures show why the use of AI alone provides an incomplete picture of the trend. A company may have implemented several AI tools without knowing whether they improve the results that matter most.

For finance and credit departments, the next step is to link each use of AI to a specific task or challenge. This could involve better prioritizing past-due receivables, reducing manual administrative work, or providing employees with a better basis for decision-making.

Collectia Group CFO Outlook 2026: The Use of AI and Demonstrating Its Value.

Better data gives the technology a clearer picture of the task

A recurring theme in London was the importance of data and an understanding of business processes. Access to a robust AI model is a starting point. How useful it becomes depends on the quality, relevance, and availability of the information it works with.

When trading loan portfolios, incomplete or inconsistent information can make it more difficult to assess the assets and price the risk. In debt collection, a lack of payment history or outdated contact information can limit understanding of individual cases.

Different tools are designed for different tasks. Predicting payment behavior, supporting a customer conversation, and analyzing a portfolio each require a clear purpose and relevant information. The choice of technology should therefore be based on the task at hand.

At Collectia, we use, among other things, AI-assisted models that help assess the likelihood of payment and support case-handling workflows. These tools help inform our employees’ decision-making, while human judgment and oversight remain an integral part of the process. This is a concrete example of combining technology with expertise in credit management.

Similar debt cases may require different next steps

In a hypothetical example, two customers owe the same amount and are past due by the same number of days. A strategy based primarily on the amount owed and the number of days past due may treat them the same way.

However, one customer usually pays on time but has unexpectedly fallen behind on a payment. The other has a long history of late payments. One responds to digital messages but rarely answers the phone. The other needs a phone call to explain their situation.

These differences may influence what the appropriate next step is. A reminder or a self-service option may be appropriate in one case. In another, a conversation with a staff member may be more helpful.

The panel discussed the trend toward a dynamic “next best action”—that is, the most appropriate next step. This approach uses relevant information to continuously select an appropriate next step as the situation evolves. The assessment is thus based on more than just a fixed customer segment. Payment history, previous contact, communication preferences, and known circumstances can all be factored into the assessment. The goal is to make each interaction more relevant and to find a realistic path to resolving the debt.

Define the success criteria before the pilot project begins

A promising demonstration is not, in and of itself, a business case. Before a company implements an AI solution, it must determine what needs to be improved and how that improvement will be measured.

When working with debt collection, relevant goals may include, among other things:

• Collection Rate: Is a larger proportion of the debt being collected?
• Cost per case: Are the total costs of processing a case decreasing?
• Processing time: Will the time required to complete a task be shorter?
• Customer Experience: Will the interaction become easier, clearer, and more helpful for the customer?

The existing process provides a baseline for comparison. The results must then be compared under sufficiently similar conditions so that changes in the nature of the cases or the customers’ situations are not mistaken for improvements attributable to AI.

Efficiency is one aspect of the assessment. Lower case-processing costs can be valuable, but so can improved customer relations or assistance in resolving cases that previously reached an impasse. A business case should include the outcomes that the solution is intended to improve, as well as the costs of implementation and ongoing operations.

Panel discussion at NPL Global 2026 in London on technology and innovation in credit management.

Integration makes the AI solution part of daily operations

Another practical issue was the gap between a pilot project and a solution that works across the organization. An AI tool may perform well in isolation but then run into problems when it needs to be integrated with customer data, telephone systems, payment systems, and existing workflows.

For example, a recommendation regarding the next point of contact becomes less useful if information about new payments or previous conversations is missing. Employees must also be able to view the recommendation in the systems they use and have a clear way to act on it.

Rolling out the solution throughout the organization therefore requires more than just choosing a model. There must be reliable connections between the systems, clear accountability for data, and a process for monitoring results after the solution goes live. These prerequisites can turn a successful pilot into a stable part of daily operations.

See also: AI is in use everywhere in Denmark. Yet governance is now the biggest obstacle

Human judgment remains central to effective debt collection

The discussion also highlighted the role of employees. Routine administrative tasks and a conversation with someone facing financial difficulties place very different demands on a debt collection agent.

Technology can support tasks that follow a familiar process and help employees prioritize what requires their attention. Complex situations, objections to a claim, and questions about what a customer can realistically afford still require thorough assessment and dialogue.

This balance must be incorporated into the workflows. Employees must understand what a tool is used for, when its suggestions should be reviewed, and how a customer can reach a human representative when needed. Proper handling of customer information and a clear division of responsibilities are part of this task.

For Collectia, the value of AI lies in supporting better credit management through both technology and experienced employees. Financial results and respectful interactions with customers go hand in hand.

The discussions at NPL Global reinforced our view on future developments: reliable data, seamless integration, measurable results, and a clear role for employees’ professional expertise. For companies looking to select a credit management partner, these are the specific issues to address in the dialogue.

Read the study behind the discussion

Read the Collectia Group CFO Outlook 2026 to gain further insights into how finance leaders are working with AI, risk, and credit management.

About the Panel

The NPL Technology & Innovation panel was moderated by Jonny Lingvall, Managing Director of Collectia in Sweden and Finland. The other participants were Sassun Mirzakhan-Saky, CTO and co-founder of Synthflow AI; Jaap Boersma, CTO at BROCC; Hardev Tumber, Group Head of Collections & Recoveries at Revolut; and Manuel Cacho García de Oteyza, Global Head of Solutions & Advisory at Accumin Intelligence.

NPL Global is organized by SmithNovak and brings together participants from the international market for investing in and managing loan portfolios. This includes, among other things, non-performing loans—also known as defaulted loans—where borrowers fail to meet their repayment obligations.

https://www.smithnovak.com/conferences/npl-global-2026.html