Advancing Analytics: FX is Embracing AI and ML

By March 16, 2023No Comments

Rising pressure to protect earnings from currency volatility is prompting FX risk managers to pursue more advanced analytics.

Sixty-nine percent: That’s the percentage of participants in a recent NeuGroup for Foreign Exchange monthly session who are already using or are exploring the adoption of advanced analytics solutions. Their reasons include improving the quality of the FX exposure forecast, conducting scenario planning and having a more refined capability to measure hedge program performance in terms of volume, effectiveness and cost.

A sweeping trend. The pursuit of more AI- and machine learning-enabled tools is not confined to the FX department. While a little late to the party, treasuries are now joining the rest of finance in looking to produce more sophisticated and actionable insight to support business and senior management decisions.

Learning ML. Most members are still in the experimentation stage, but according to one participant in the session, the only way to come up the learning curve is by doing.

  • “We’ve been on a journey,” the FX manager said. One area of focus has been enhancing the accuracy of the forecast using machine learning algorithms. “Forecasting is a challenge,” he said, “because our cash flow forecasting process involves reaching out to over 150 entities worldwide.”
  • This corporate has been helped by the establishment of a finance center of excellence (COE) for analytics, staffed by data scientists who can write and update algorithms. “We are using data scientists to do more of a top-down format,” he explained. “We added the data scientists internally to help with this.”
  • While the company has had some success in forecasting net exposure, “We still don’t have visibility to explain variances.” With pressure from management, getting to the root cause is essential.
  • “We’re not perfect today,” he admitted. “We need to update the algorithms so we can get more details and set new parameters. We are making progress in improving accuracy but also automating part of the job to free up staff capacity.”

Leveraging embedded tools. Like the company above, other treasuries are taking advantage of finance-wide automation initiatives to gain access to better tools, for example through the implementation of SAP S/4HANA.

  • “We have been on a journey almost for almost five years to integrate all of treasury into SAP, initially using SAP Treasury, but now in a migration to S/4HANA,” one member shared.
  • Having an all-in-one solution was a big draw. “It has allowed us to get rid of third parties and give us more direct information for reporting,” she added. The third parties include vendors of FX risk management solutions.
  • “Our first step is to aggregate and clean the data, so we can all pull from the same source,” she said. “Data analytics is huge for us, and we’re working to build those models.”
  • Another member who is implementing S/4HANA said that the immediate benefit for treasury has been real-time visibility into data. “We’re very much in the limelight right now because of the extreme volatility and having this technology has helped us tremendously.”
  • “Once [SAP] gets full functionality in the platform, they will give other vendors a run for their money,” the participant predicted.

Three key considerations. The migration to SAP S/4HANA is a trend across NeuGroup finance and treasury peer groups. For the company referenced above, the transition was relatively smooth because it was on a single instance of SAP from the start. The FX manager acknowledges the migration is more challenging for organizations with multiple ERPs or even multiple instances of the same ERP.

Here are three things to consider, as the pressure to improve analytics mounts:

  1. The acceleration of finance digital transformation is bringing new opportunities for treasury teams that do not require securing a dedicated budget. With cost pressures mounting, it’s a lot easier to build a business case for benefiting from in-flight implementations.
  2. At organizations with a fractured system environment, treasury should consider whether a TMS vendor is adding or planning to add AI-enabled analytics, for example for forecasting. Another option is to adopt a standalone analytics solution to speed up the development of modeling capabilities; some tools may already be in use in another part of finance, e.g., FP&A.
  3. Finally, if treasury and finance are looking to build analytics core competency, there are ways to leverage a COE model to amplify the reach of a small group of data scientists. The benefit is the ability to try out different use cases and improve on the go.
Justin Jones

Author Justin Jones

More posts by Justin Jones