CUSTOMERS PARTNERS RESOURCES  
 
AutoClass
 

An Artificial Intelligence (AI)
based commodity classification
software

Analytics
 

Provides visibility into an agency's spend & supplier-base

Spend Miner
 

Identifies savings opportunities by mining through item-level data

ETL
 

Data extraction & aggregation

   
 
Request for information
 
 
    AutoClass
 

Overview
AutoClass is an Artificial Intelligence (AI) based commodity classification software that is designed to read "Item Descriptions" from purchase transaction and associated supplier names, and automatically assign a granular (8 digit) UNSPSC code to the same.

AutoClass is designed to process real world spend data which contains a lot of acronyms, cryptic characters and sometimes even incomplete descriptions. Some transaction data may contain both supplier names and item descriptions, while others include only supplier names with associated GL codes.

AutoClass runs in a completely hands-free environment, integrated with Data Warehouses (from vendors such as SAP, Teradata, Oracle etc) or plugs-in to data-marts built in-house by federal agencies. Using patented AI methods, AutoClass provides far superior classification results in terms of consistency and accuracy, as compared to alternate methods of 'mapping' material group codes assigned by buyers to UNSPSC.

Real-world spend data is inaccurately described and categorized in most transaction systems, because of the lack of systems and processes to capture good quality data at source. This necessitates the use of AI technology for accurate re-classification, prior to performing spend analysis.

Advantages

  • Automated Approach - Provides consistency in classification, higher throughput and classification accuracy.
  • AI based technology - Provides for a closed loop learning mechanism, which improves the tool's performance over a period of time in the form of increased classification accuracy.
  • Accuracy - AutoClass is capable of classifying short and poor descriptions, supplier name abbreviations, wrongly spelt words, which are the characteristics of real world commodity and master data.
  • Language and Taxonomy independent - It can be used for classifying into any schema like UNSPSC, eClass, NATO or even to customer's custom schema and it can classify data in any language. It can be trained in any industry vertical.
  • Task Scheduling - This is a key feature for classification on varying time tables. With an easy-to-use GUI, AutoClass can be scheduled to access files from designated Internet/Intranet locations via FTP.
  • Real-time classification - AutoClass Live Connect, the real-time classification module of the core product can be integrated with existing eProcurement or purchasing systems for classification and coding of transaction records right at source.

 I want to see a demo

 AutoClass Data Sheet (234 Kb)

 
Illustration for data classification (before/after)
Vendor Rationalization
(Click for enlarged view)
Spend by Supplier
(Click for enlarged view)
 
UNSPSC Category
(Click for enlarged view)
 
 

 Tell me more about AutoClass

 
   Acquisition Spend Analysis Solution Components
  Zycus Software Tools/Products to enable automated repeatable Spend Analysis
Autoclass

An Artificial Intelligence (AI) based commodity classification software
Read More

 
ETL

Data extraction & aggregation
Read More

Analytics

Provides visibility into an agency's spend & supplier-base
Read More

 
Spend Miner

Identifies savings opportunities by mining through item-level data
Read More

 
 
 
UNSPSC for better
Spend Analysis
An insightful 40-page booklet on adopting global taxonomies for better Spend Analysis
Download
Spend Analysis-
Challenges for
Government Agencies
The Spend Analysis imperative for federal agencies - and how to do it right the first time, every time..
Download
Other Resources

WHITE PAPERS

CASE STUDIES

ARTICLES
PRODUCT DATA SHEETS
Register for live demo
Site best view in 1024x768 resolution | © 2006 Zycus Inc. | Send comments and suggestions to webmaster@zycus.com