Data Cleansing & Categorization

Use Artificial Intelligence in procurement

Classify records, suppliers, and materials automatically

DataCategorizer automatically classifies transactional data such as invoices and orders as well as master data such as suppliers. By classifying data thoroughly and granularly based on a company-specific or standard taxonomy (e.g. eCl@ss, UNSPSC), you bring unprecedented transparency to strategic procurement while creating more accurate spend controls and informative analyses.

DataCategorizer builds on your existing ERP or data warehouse systems. At Orpheus, we refine your pool of information and boost its quality significantly. Through this higher transparency, you lay the foundation for higher savings and reliable procurement KPIs.

DataCategorizer - Classify records, suppliers, and materials automatically
DataCategorizer - Classify records, suppliers, and materials automatically

Identify bundling and other savings potential and derive trends through classification

  • Classify invoices, orders, or suppliers effectively (high automation)
  • Work with your choice of taxonomies (eCl@ss, UNSPSC, company-specific)
  • Apply state-of-the-art classification methods (rule-based methods, expert system) and statistical methods based on word frequency or text mining
  • Create highly granular classifications (depending on data quality)
  • Enhance your data and its quality
  • Use DataCategorizer as an add-on for your data warehouse (e.g. SAP BW/BI or SAP HANA)

DataCategorizer enables in-depth spend analytics:

  • Calculate your spend volume and prices by product and sourcing categories (e.g. in line with eCl@ss or UNSPSC taxonomy)
  • Identify which suppliers serve several categories and align your sourcing strategies accordingly
  • and much more!

Many procurement professionals still work with BI analyses that are misleading or lack the necessary details to interpret the facts appropriately. Thorough, highly automated classification is essential to create a realistic view of your complete spend.

DataCategorizer - Classify records, suppliers, and materials automatically

Benefits for your business

What makes DataCategorizer different than other classification solutions in strategic procurement?

Easy and intuitive to use

Orpheus makes it easy to optimize your procurement through advanced semantic data management. With the help of our intuitive templates, users can define rules and configurations based on their classification knowledge and apply them to spend data (e.g. invoices, orders, suppliers).

Higher automation through artificial intelligence

Good software should help automate routine tasks and avoid errors. Based on this guiding principle, Orpheus has developed an application that optimally supports you in your work. As a procurement professional, only you know how much automation makes sense for the specific task at hand! Use text mining and other AI methods to create more transparency in procurement.

Support for various taxonomies and classification standards

You can use DataCategorizer no matter if you use eCl@ss, UNSPSC, or a custom taxonomy developed for your company’s unique structures.

Higher classification quality through external data

Linking or enriching spend data with additional information is as simple as uploading an Excel sheet or other external files to DataCategorizer. You can also automatically download information (e.g. from suppliers) from the internet.

Flexible combination of manual and automatic steps

DataCategorizer delivers the automated, intelligent tools to flexibly support users in the classification process. Users can choose the level of automation that best meets their requirements.

Easy-to-use classification rules, easy-to-analyze results

DataCategorizer allows you to test each individual rule and suggested classification and simulate the effects before it goes into productive use. You can also analyze the final categorization results, for example, in order to understand exactly how a certain rule set affected the data.

Classifications in the depth and detail you need

Using nested categorization trees, you can execute classifications in the granularity you need. More in-depth, complex classifications, however, also require higher quality data and more sophisticated categorization methods. If your data quality has not yet reached that level of quality, classification methods based on text recognition and comparisons (i.e. text mining) are better than rule-based ones. Either way, DataCategorizer has both types built-in.

Automatic recognition of data inconsistencies and classification errors

Many spend data errors, including incorrectly assigned material groups or general ledger accounts, stem from the ERP systems themselves. These types of false information construe your data and can lead to classification problems. DataCategorizer recognizes these inconsistent or false data constellations and cleans the data before classification.

Intelligent classification algorithms

Algorithms, simply known as methods, are the heart of any classification system. Over the years, Orpheus has built vast intelligence into these methods to produce highly automated results that you as a user understand. At Orpheus, there is no automation without controls. Aside from rule-based algorithms, we rely on text-mining methods from the field of artificial intelligence. But you don’t have to worry about that. DataCategorizer automatically applies the correct method depending on the data at hand.

Visualizations for simpler spend data classifications

No computer can replace a person when it comes to identifying patterns and making classifications! That’s why DataCategorizer offers vast visualization options so that you can truly “see” and interpret your data.

Better, more granular classifications – step by step

The best way to achieve highly granular, error-free categorization is to break it down into a series of steps. Start by classifying materials, general ledger accounts, and suppliers separately. Afterwards, you can generate more complex rules with several attributes (e.g. supplier/GL account rules or keyword rules). DataCategorizer offers many different options with up to 12 rule types.

Connect with our experts