How artificial intelligence and machine learning are transforming Treasury

artificial intelligence and machine learning

Developments in artificial intelligence (AI) and machine learning (ML)

Nils A. Bothe

Partner, Financial Services, Finance and Treasury Management

KPMG AG Wirtschaftsprüfungsgesellschaft


Related content

Artificial Intelligence


The digital revolution with its origins in the 80s made significant strides in the areas of computation and user experience, but there so much more to come! For instance, developments in artificial intelligence (AI) and machine learning (ML) offer companies possibilities never seen before to gain valuable insights and improve their performance.

For corporate treasurers, these technologies will bring advantages, such as an improved accuracy of liquidity forecasts, automated reconciliation of accounts and optimized hedge ratios. 

So how does this affect the everyday life of a corporate treasurer? Will human treasurers soon be replaced by robots? One thing is for sure: artificial intelligence (AI) and machine learning (ML) will change and improve treasury operations.

The AI/ML revolution

Going from Deep Blue to AlphaZero

Looking at computer chess over the last 20 years is a good way to see how powerful AI has become and to show how much progress has been made in this area. In 1997, supercomputer Deep Blue wrote history by beating world champion Garry Kasparov in a six-game match. Then in 2005, the freestyle chess championship (featuring grand masters and a combination of humans and computers) was won by amateurs who worked with computers. Their performance showed how the symbiotic relationship of humans and machines can improve the skills of amateurs.

In 2018, Stockfish (a previously unbeatable chess program) was defeated by AlphaZero, a deep learning algorithm that taught itself to play chess in 4 hours. Of the 100 games played, 72 ended in a draw, with AlphaZero winning the remaining 28 games. An analysis of the unconventional techniques applied by AlphaZero following the game could inform the strategies of future chess champions.

Unimagined possibilities

Chess is just one of the many possible applications of AI. The example dating back to 2005 shows how using deep learning brings advantages that go beyond what humans or computers can accomplish individually. The development of new ways of collaborating between humans and AI in the digital age will without doubt change the way certain economic issues will be tackled and solved going forward.

Companies can use artificial intelligence to discover new forms of value creation involving technology, improve performance and increase the value of their company.

Treasury teams can benefit from artificial intelligence and machine learning

AI and ML could make treasury teams more effective, allowing them to add value to the company they serve. There are many ways how AI can be used in the context of treasury. 

A possible use of AI might be in risk management. Using AI to strengthen controls could improve security through a better recognition of anomalies and countermeasures to combat fraud. By using the data available in the TMS, ML algorithms can spot and analyze patterns or anomalies and gain insights that would be difficult for the human eye to spot. This information may then be used to take targeted and informed decisions, thus improving the value of Treasury to the company. Companies may achieve financial gains with AI, improve their preparation and planning as well as validate data. 

Moreover, these tools can free up the Treasury team’s time by performing tedious everyday tasks so that they can become more strategic actors and work more efficiently. AI can take on the work of treasurers and often perform it in a fraction of the time originally used for it. 

Using machine learning for cash forecasting

For most companies, cash forecasting is of the first priority. Human beings are unable to spot trends and patterns in large amounts of data. Machine learning can help navigate all this information to automate and improve the forecasts.

Many internal and external events could cause Treasury to change its cash forecasts at very short notice. Such events could be the merger of two companies, an unexpected rise or sharp drop in sales, an unforeseeable event or a large overhang in stock for the coming months. While the computer provides a cash forecast based on historical data, the Treasury team will have to communicate with the computer at one point to add insider and current daily data in order to get the best results for the forecast. 

ML can accelerate the forecasting process. Instead of having staff at subsidiaries prepare a manual (and most likely skewed) forecast, AI will generate the forecast for the entire group in seconds based on the data. AI is not only quicker but also more precise. Beyond that, companies may also use ML algorithms to validate cash forecasts and to save costs when implementing it. 

But this is just an example. There are many other use cases that would drastically reduce the implementation costs and the go-to-market time of a TMS.

But what about the human being sitting in front of the computer?

Tools such as machine learning and artificial intelligence will not replace human treasurers. Treasury experts have to understand that their role will remain a strategic one for the company, even if AI or ML is being used going forward. Using AI simplifies the analysis of data therefore making it easier to understand. This makes for more informed decisions, for instance in payment operations or hedging. It also allows developing new strategies, supported by the computer. Humans should be less involved in repetitive and manual tasks which can be supported by AI so that they can use their time for more precise insights into patterns and risks that would be undetectable without technology. In the end, it will be a combination of computers and humans that will make a team successful. Humans will always hold more company-internal data and knowledge than a computer, which will influence and support strategic processes and decisions. Humans will always be needed to take decisions. However, these decisions will be made on a more solid information base, with more analyzed data and faster response times.

By invitation; written by:
Viola Hechl-Schmied
Product Owner
ION Treasury

Source: KPMG Corporate Treasury News, Edition 105, October 2020

© 2021 KPMG AG Wirtschaftsprüfungsgesellschaft, eine Aktiengesellschaft nach deutschem Recht und ein Mitglied der globalen KPMG-Organisation unabhängiger Mitgliedsfirmen, die KPMG International Limited, einer Private English Company Limited by Guarantee, angeschlossen sind. Alle Rechte vorbehalten. Für weitere Einzelheiten über die Struktur der globalen Organisation von KPMG besuchen Sie bitte

Connect with us


Want to do business with KPMG?


loading image Request for proposal

Stay up to date with what matters to you

Gain access to personalized content based on your interests by signing up today

Sign up today