ElectrifAi
July 8, 2021

Veolia Gains Visibility & Control Over Multiple Data Sources

Veolia North America sought out ElectrifAi in September 2018 to optimize spend management problems with machine learning capabilities. Their objective is to have a simple and flexible solution that allows category managers to do spend analysis with accurate data classification.

Veolia group is the leader in optimized resource management. The group designs and provides water, waste, and energy management solutions that contribute to the sustainable development of communities and industries. They are uniquely positioned to provide solutions in water, waste, and energy that promote sustainability and the circular economy.[1]

Challenges

ElectrifAi was asked to solve Veolia’s supplier mix by category; upgrade inefficient legacy systems used for reporting as the systems did not include spend data from all of Veolia’s entities; many of their spend records lacked data, making spend categorization a challenge.

Solution

Implementing ElectrifAi’s SpendAi offers flexibility and accurate visibility across Veolia’s multiple data sources and combines it into a single view. From there, the data was classified in a manner that fits the need of a cross-functional team.

ElectrifAi’s SpendAi beings with data cleansing using three machine learning models: Company Name Standardization, Vendor Name Grouping, Transaction Classification. Clean data is then ported to a highly configurable SaaS product for analysis.

These machine learning models are also available on Amazon SageMaker Marketplace (AWS) and provide easy access to the models.

Company Name Standardization

Leveraging 1st and 3rd party data, standardize and cleanse the common problem of company name variability. Regardless of acronyms, acquisitions, or other company evolutions, this algorithm will create a uniform naming convention for company names in your data.

Vendor Name Grouping

Reduce the manual effort to clean and group company names by identifying which companies are subsidiaries of other companies. Technical highlights include applying Convolutional Neural Networks (CNN), fuzzy matching techniques, and key collision. Using the data input of procurement transaction records, the model outputs a suggested group name for each vendor name.

Transaction Classification

Each procurement transaction is automatically classified into a standard taxonomy using company description and line of business. The model classifies 95+% of total spend. Technical highlights include applying a Convolutional Neural Network (CNN) classification model. Using the data input of procurement transaction records, the model outputs suggested transaction categories.

Conclusion

Veolia North America is on a great path towards spend management optimization. From categorized suppliers, legacy systems organized and improved, and spend categorization optimized, Veolia’s Procurement team is now more efficient and working towards the company’s goals for growth and customer satisfaction.

Do you have concerns about your spend management? ElectrifAi is here to help with deep domain expertise, experienced data scientists, and machine learning models customizable to your specific business problems. Contact us today and find out how to optimize your spend!

[1] https://www.veolianorthamerica.com/what-we-do?page=1

You have data, we have solutions!

Find out what ElectrifAi has to offer by filling out the information below.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.