人工智能驱动的金融科技：使用 DATABRICKS 和 MLFLOW 进行 AR 预测.pdf

1、2024 Databricks Inc.All rights reserved1AI-POWERED FINTECH:AR FORECASTING WITH DATABRICKS&MLFLOWFloriant Sturm&Julie VanackereFloriant Sturm&Julie Vanackere12/06/202412/06/20242024 Databricks Inc.All rights reservedJulie VanackereJulie VanackereDataData scientistscientistNice to meet you!Nice to mee

2、t you!2daidaloFloriant SturmFloriant SturmCoCo-founderfounder2024 Databricks Inc.All rights reservedOutlineOutline2 main topics2 main topics1.Why Accounts receivable(AR)forecasting?2.How do we approach this?3.How can AR forecasting become a plug&play solution and what are the technical requirementst

3、echnical requirements?Technical perspective1.Automated infra deployments2.Standardized feature engineering3.Standardized ML training4.An automated ML lifecycle5.Monitoring for customer confidenceBusiness perspective2024 Databricks Inc.All rights reserved2024 Databricks Inc.All rights reserved4Busine

4、ss Business perspectiveperspective2024 Databricks Inc.All rights reservedPredicting when a customer will pay their invoice in the future This will help you to anticipateWho will pay lateHow much cash is to be expected in the next x periodContact strategy:incentivize customers that are likely to pay

5、late=proactive approach Currently companies use a reactiveapproach to chase“late-payers”They contact them after it is too lateThey know their outstanding invoicesBig issue“Cash flow is the pulse of the company”Why did we focus on AR forecasting?Why did we focus on AR forecasting?5What is AR forecast

6、ing?What is AR forecasting?AccountsAccounts-Receivable Receivable Outstanding invoicesForecastingForecasting2024 Databricks Inc.All rights reservedQuick win!We only need the historical invoice data to get startedPredicting future cash flows reduces:Credit riskPlan expenses,investments&potential savi