Artificial Intelligence (AI) in cash collection

Artificial Intelligence, the main topic of the decade that began in 2020, is transforming many fields, raising questions, challenging us, and even making us question our very condition as human beings. Are we still useful, or is AI, or will it become, more powerful, condemning us to irreversible obsolescence?

The idea of ​​humans being overtaken by machines, a popular theme for fantasy and science fiction authors for decades, has become, in just a few years, a dizzyingly real and concrete issue. However, Artificial Intelligence is not a recent invention. Its beginnings can be traced back to 1956, 70 years ago, with concrete applications that remained limited to well-defined processes. However, the exponential increase in the computing power of CPUs (processors) and GPUs (graphics cards), the progress of machine learning models, the exploitation of massive datasets available on the internet, and the development of highly innovative algorithms have enabled the emergence of generative AI, which represents a fundamental shift. The application areas of this AI are so vast, and the potential for revenue and power so enormous, that they are attracting colossal investments, amounting to hundreds of billions of dollars. By 2026, global investment in AI is estimated to exceed $2 trillion, four times the French national budget.

Everything is happening very fast. AI capabilities are improving daily to such an extent that their evolution seems unpredictable and appears to be beyond the control of any human "authority." Consequently, many intellectuals, politicians, entrepreneurs, and even AI specialists regularly publish manifestos to warn of the dangers of AI and to halt, at least temporarily, research and advancements in this revolutionary technology. However legitimate these efforts may seem, one can reasonably doubt their ability to achieve their objective, given the powerful factors driving AI.

AI is therefore very much present and is also transforming credit management roles across all its aspects:

• Customer Risk Management: improved risk assessment based on a more comprehensive and in-depth range of data, advanced scoring, and increased predictability of default risk.
• Dispute Management: AI-assisted qualification of customer feedback for more effective handling of customer dissatisfaction.
• Cash Collection: more tailored recovery actions for each customer, more effective customer interactions leading to significant performance improvements.
• Performance Assessment: in-depth analyses and near real-time reports, based on more complete and complex data, enabling much greater predictive and analytical capabilities, linked to actions proposed.

In credit management, the relevance of AI lies in its incredible ability to imitate human behavior, or even to outperform it in certain aspects, on a mass of data inaccessible to a human brain. Its limitations lie precisely in the fact that it is not human, and that it is completely foreign to purely human concepts such as sensitivity, intuition, empathy, imagination, foresight, etc. On these essential aspects, AI can only give the appearance and the illusion that it incorporates them. This is not the case. It's just a computer program, however impressive it may be.

Artificial Intelligence in Cash Collection

Cash collection is a profession that integrates many diverse and seemingly contradictory concepts: the necessary orthodoxy and rigor of accounts receivable, financial challenges aimed at preserving profitability and improving cash flow, customer relationship quality, contributing to customer satisfaction through the efficient handling of disputes, supporting revenue growth, etc.

This complex mix of challenges and multiple required skills represents a particularly favorable environment for Artificial Intelligence to truly be what it is: a super assistant.

With a vast quantity and variety of data, AI thrives where the human brain is slower and has to consider it one by one. With a considerable number of clients to manage, involving a multitude of internal and external exchanges, AI synthesizes hundreds of these per minute, qualifies them, and proposes follow-up actions, whereas the human brain processes one element at a time.

The relevance of AI is determined by the data it has access to. This results in two essential principles:

• Interconnection between systems: This allows for the combination of multiple complementary pieces of information that AI can leverage to increase its performance and accuracy. AI for data recovery is useful if, and only if, it is fed with accurate and comprehensive data. This is one of the reasons for the constellation of connectors in My DSO Manager, a cash collection software, which allows for the integration of numerous data points beyond just accounting data.


• The real-time dimension.
Along with data interconnection, the real-time or near-real-time dimension is crucial for several reasons:
• It allows AI to be fed with data that is always up-to-date.
• Real-time recovery software enables the implementation and exploitation of the efficiency of AI, which is itself highly responsive and operates in real time.
• Performance assessment: in-depth analyses and near-real-time reports, based on more complete and complex data, allow for much deeper predictive and analytical capabilities, linked to proposed actions.

Interconnectivity, real-time processing, and agility are the key words that go hand in hand with Artificial Intelligence.

The Use of Artificial Intelligence in Cash Collection

The applications of AI in cash collection are numerous, and concern both customer interactions and the analysis of payment behavior, as well as the implementation of appropriate recovery strategies.

• AI for Managing Customer Interactions

  Cash collection software allows you to effectively follow up with customers and capture their responses. In this respect, AI excels, particularly given the volume of customers and the variety of their feedback. Whether they respond via a portal (an approved platform for electronic invoicing), an interactive page, by email, or by phone, AI synthesizes, rephrases, categorizes, and qualifies customer feedback to provide an appropriate response and adapt the follow-up strategy.
  
  Whether for handling a dispute, monitoring a settlement agreement, or any other situation, AI manages customer feedback for efficient and rapid processing, regardless of the volume.
  
  

  Example with MAIA Feedbacks

  MAIA is the generative AI of My DSO Manager, and intervenes at several levels within the software: user support via MAIA Chatbot, prioritization of actions with MAIA Actions, and qualification of customer feedback with MAIA Feedbacks. Each customer response, whether received via the interactive page or by email, is analyzed by MAIA, which then assigns a status (e.g., administrative dispute), provides feedback phrased like that of a cash collector, and suggests the most appropriate next step.
  
  
  
  
  This example perfectly illustrates the need to combine AI with a support system, i.e., software that includes suitable functionalities. The interactive emails included in My DSO Manager allow you to communicate effectively with customers about accounts receivable management. They can reply from their portal, and the AI ​​will then interpret their feedback and act accordingly.
  
  
  AI is not a solution in itself; it is part of a solution that also includes systems, the interconnection between systems, and the interconnection between people. It is only the combination of these elements that is relevant and that makes it truly effective.



• AI to Define the Most Adapted Collection Strategies.

  Imagine you're a cash collector with only one client to manage. This client would be monitored very closely and optimally. Every follow-up action would be taken at the right time, with the right people, and perfectly tailored to the situation, with the goal of getting paid and improving customer satisfaction. However, when you have not just one client to manage, but hundreds, probably thousands, this highly personalized approach is humanly impossible to implement.
  
  This is where AI comes in, applying this level of monitoring to thousands of clients simultaneously, adapting recovery strategies, just as you would, to the specific needs of each one.
  
  To do this, it's essential to teach the AI ​​your methods and best practices. AI doesn't operate in a vacuum; it needs to be trained to apply what you would do in its place, if you had the opportunity, based on your company's activity, its customers, its business culture, its credit management practices, etc.
  
  

  Example with My DSO Manager's Search & Assign AI:

  The Search & Assign AI allows, based on all the information and characteristics of customer accounts, the dynamic and near real-time assignment of the right recovery strategies to each customer. It relies on all the information present in the platform: customer type, risk score, customer account status (payment delays?), payer profiles, etc.
  
  Systematic adaptation allows for the automation of collection actions without any loss of quality; quite the opposite, in fact. Collection agents oversee actions carried out at the right time and with the highest quality. Not only do they perform ten or a hundred times more actions than before, but these actions are of higher quality and are always completed on time.
  
  
  
  
  
  This use of AI is an excellent example of how to approach it correctly and of the complementarity between humans and technology. Humans determine the direction, AI implements it efficiently and flexibly.
  

• AI to Prioritize Collection Actions

  Artificial intelligence allows for a much more effective automation of collection actions. However, it doesn't have all the information related to the business relationship with each client. In some cases, particularly with high-stakes clients or situations, it's not enough. A truly intelligent AI is therefore capable of determining when it shouldn't act, but when qualified human intervention is necessary.
  
  In these cases, manual or semi-automated actions are essential and are carried out by collection agents. AI helps prioritize collection actions based on the stakes involved. It is capable of taking into account a multitude of criteria: the amount of the delay, the client's usual payment behavior, assessment of the risk of insolvency, the age of the debt, etc., in order to guide cash collectors in prioritizing the actions with the highest stakes, thereby preserving cash flow and profitability.
  
  

  Example with MAIA Actions (under development):

  
  
  
  Cash collectors are thus encouraged to focus where their expertise and knowledge are needed and to prioritize the highest-stakes actions.



• AI to anticipate and predict

  Time is one of the keys to effective cash collection. Taking timely follow-up actions has always been essential for success. This principle is easier to adhere to with the use of cash collection software that integrates tailored business functionalities.
  
  Artificial Intelligence, however, allows us to go further, particularly in predicting cash receipts, for optimized cash flow management. Thanks to all the information contained in the cash collection software, including past transaction history, the current status of customer accounts and each invoice, AI can generate much more accurate cash flow forecasts. Combined with real-time software that allows for the instant consolidation of all future cash flow forecasts at the corporate level (for groups), AI provides added value that goes beyond accounts receivable management and improves overall cash flow management.
  
  

  Example with the My DSO Manager cash flow forecast report:

  
  
  This report is frequently integrated into cash management software to improve the forecasting of future cash flows.



• AI for Comprehensive Analysis

  The sheer volume of customers and transactions has always posed a challenge for credit managers. digitalization and the interconnection between systems only increase the quantity and variety of available data. Without the right tools, effectively leveraging all this data becomes increasingly difficult for a human or a team.
  
  However, Generative Artificial Intelligence (GAI) is inherently very efficient at extracting key insights from large datasets based on a defined framework. In just a few seconds, an Excel file containing a large amount of data can be analyzed by Mistral or ChatGPT.
  
  In credit management and cash collection, it is strongly advised against sharing your company's accounting data with these systems for security and confidentiality reasons. Once again, cash collection software like My DSO Manager possesses all the relevant data and offers integrated AI that guarantees complete data confidentiality.
  
  MAIA Report (in development for 2026) will allow users to analyze a client, a report, or a platform from any desired perspective, in order to obtain recommendations on customer risk management or cash collection. Performance assessment and guidance for improvement will be even more readily available.