AppZen's Artificial Intelligence

Follow

 

Screen_Shot_2019-06-27_at_9.40.14_AM.png

AppZen's artificial intelligence specializes in 4 key areas:

 

  1. Data Extraction and Understanding
  2. Data Enrichment and Augmentation
  3. Risk Assessment
  4. Behavioral Tracking

All expense reports are ingested and processed with AI. AppZen Expense Audit evaluates 100% of expense lines and receipts within a report and then assigns a risk score of low, medium or high risk.

High-Risk - represents expense lines that require further review from an auditor. These expenses have been detected as a potential user error, compliance violation, or fraud.

Medium-Risk - represents expense lines that are often seen as an accepted business risk. For example, an auditor will usually approve a report where an employee exceeds the daily meal limit by $5 or $10. Even though these types of business risks infringe on company policy, they often do not warrant the review of an auditor as they are typically accepted and approved. Therefore, AppZen would recommend a company to automatically approve medium-risk expenses.

Low-Risk - represent expense lines that are compliant with company policy and exhibit no characteristics of fraud. These expenses should be automatically approved.

AppZen logs all risk data and compares it across all employees, assigning each employee an AppZen Behavioral Index (ABI) score. This rating provides an analytical way to compare risk among an employee base. If an employee demonstrates a repetitive behavior of submitting an excessive number of medium-risk or high-risk expense reports, AppZen will flag this individual as a risky employee shown through a higher ABI score.

An auditor can review an employee's ABI score and gain a more comprehensive understanding of an employee's expense spend behavior by viewing the Auditor Dashboard.

AppZen uses a combination of deep learning, machine learning, and semantic modeling as part of AppZen's AI.

0 out of 0 found this helpful

Comments

0 comments

Article is closed for comments.