Consider the time and money needed to put AI to use in cash forecasts.
The pain of accurate cash flow forecasting is all too familiar to treasurers used to struggling with multiple models, getting access to data and convincing business units of the importance of timely and accurate forecast submissions. No wonder many treasurers are eager to learn more about using artificial intelligence (AI) and machine learning to make the job easier.
Good news, bad news. The treasurer of a large multinational gave his peers plenty to think about at a recent meeting by telling them the encouraging news that his company has automated the process of cash flow forecasting successfully. The bad news: It took a decade.
- It took 10 years, he explained, because that’s how long it took the company to create and use a single instance of SAP. “One instance is necessary for an automated cash forecasting process,” he said. “It was a significant project.”
Single truth source. “Our forecasting uses machine learning and you can only do that with one data hub that is a single source of the truth,” the treasurer said, adding that senior management must mandate that everyone uses the hub, not just treasury.
- The universal data hub sits on top of SAP and takes information directly from Quantum and other systems, including those for budgeting, tax and logistics. Spotfire, from TIBCO, sits on top of the data hub for reporting.
Share the financial pain. The treasurer said getting budget for the project was a big issue, but commercial groups shared in the cost, which was well into the millions of dollars.
Worth it. “The benefits significantly outweighed the cost given the ability to proactively identify where pools of cash would be created allowing early action to dollarize and repatriate excess cash,” the treasurer said.
