For the fourth time, DACT has organised a competition to identify the most outstanding treasury project in 2023. Following an exciting competition, we are delighted to unveil the recipient of the DACT Treasury Award 2023. With four noteworthy nominations this year, the ASML project ‘Improved FX Exposure Forecasting with Artificial Intelligence’ emerged as the winner, securing the highest number of votes from DACT members. If you’re intrigued to learn more about the winning project, the ASML team is set to present it at the Treasury Targets Meeting scheduled for 7 February 2024 at van der Valk Breukelen.
About the project
In 2023, ASML implemented a new AI powered material intake forecast model to increase the effectiveness and efficiency of its purchase FX hedging program.
ASML sources components for its chip machines from a large worldwide network of suppliers. The vast majority is sourced in Europe but some of them are sourced from the United States and paid in US dollars. ASML sells its systems in euros and is therefore exposed to EUR/USD exchange rate fluctuations. To mitigate the effect on its P&L, there is a purchase FX hedging program in place.
The fundament of this hedging program is the forecast of expected US dollar-denominated material intake. Prior to implementing the AI-powered forecast, the accuracy of the forecast was low (70%), it was manual and labor intensive, Excel-heavy, time-consuming and prone to error. Therefore, this project was launched to find a better solution.
The new forecast is generated by a fully automated AI Machine Learning model which was developed internally. The AI model recognizes patterns and trends in historic actuals and uses those to predict future intake. With every new month, the model becomes ‘smarter’ and improves the forecast accuracy even further. As a result, the forecast accuracy went up from 70% to 96%, making the hedging program more effective.
The project is a notable example of a close collaboration between two ASML teams – Treasury and Data Science. The expertise of both teams was required to create a working solution.
ASML decided to share this solution with the broader Treasury community to show that AI applications can be used and developed broadly in any company, big or small. This particular AI model is based on a free open-source Python algorithm that was further optimized to ASML’s use case. This approach makes AI solutions easily accessible, scalable, and widely usable for any person or company. AI also helps to improve ASML’s understanding of the data while decreasing time spent on manual data gathering and processing, thereby freeing up time for more value-adding activities.
Key business benefits
- Improved exposure forecast accuracy (from 70% to 96%).
- More effective hedging, further reducing the impact of EUR/USD movements on ASML’s P&L.
- More efficient input for hedging without waiting time and manual effort/errors.
- The AI solution is learning every month, ensuring adaptability to business developments.
- Resources are freed up for more value-adding activities instead of data processing.
- Increased insights into the underlying data.
The experience gained in this project will be applied to many more use cases within ASML.
Award ceremony
Award ceremony will take place during the ‘DACT Treasury Targets Meeting 2024’ on Wednesday, February 7th. DACT members and management representing the winning Treasury projects will be invited to attend.
About DACT
DACT, the Dutch Association of Corporate Treasurers, is the association for corporate treasurers and treasury professionals in the Netherlands. The association has more than 730 members, working at multinationals and large and medium-sized companies as well as government and non-profit organisations. DACT promotes the professional development of its members and the treasury profession, both within and outside the association.
