Natural Language Generation Made Easy with Alteryx
With over 5 million private and business clients, this insurance company is one of the largest financial service providers in the Netherlands. It offers a complete range of products including insurance, pensions, savings, mortgages, investments and loans. Learn in this use case how their finance department used Alteryx to execute their Natural language generation (NLG) project to transform financial figures into analytical reports.
“Every month our finance colleagues need to write an executive summary of the monthly operating results. They need to download dozens of Excel files from our ERP system, look into the correct cell for the figures, then analyze the figures and share some highlights with the management”, Sufeng Li, Robotics Product Owner.
Sufeng and his colleagues were spending 4 working days every month to prepare the executives summary. After analyzing the figures, they created visuals on PowerPoint that had 50 slides. Sufeng has been using Alteryx for the past 2 years and the finance department has 22 Alteryx users. Their goal was to make their computers read, decipher, and understand the financial figures in the Excel spreadsheets to make sense of the human languages in a manner that is valuable.
- We used iteration macros to read all the input files with different schemes. One input is the report template and the other input is data. We named the macro Ara, a type of parrot.
- Data cleansing and merge to Excel
- Searching for the target data using the labels defined in the input template
-
Generated the analysis to Excel. The output are sentences
-
Visualize the data in Power BI
Describe the benefits you have achieved
- For the PoC project, one-week workload is saved by this tool. This time-saving happens during the month closing time-critical period.
- Potential saving can be foreseen by applying this tool on other similar reports.
