Delivering Next Level Data Validation

Automation has proven to be a key strategy to drive down the cost and efforts involved in complex migration projects. When the world’s largest social media company found themselves in the middle of a large ERP migration project, they figured it was time to leverage RPA investments to ease the pain of the data migration and validation processes. The goal was to minimize human errors while seamlessly moving data across platforms and expediting the migration of on-premise Oracle EBS data to Oracle cloud.

When migrating to Oracle Fusion Cloud, data conversion can become a significant challenge. Numerous data tables will need to be reformatted, and employees would need to populate items based on the differences with the sources systems.

No matter the source systems, challenges will inevitably emerge during the data validation process. Some of the common challenges organizations face during data validation are that data tables are often outdated. The amount of data rows inside a file can reach tens of thousands. Additionally, a significant challenge is that conversion data must be normalized; this means that data needs to be extracted, reviewed, any missing data must be populated, validated, and even reformatted before conversion. During cloud migrations, the data validation process is very high-volume, manual and tedious—this where RPA plays a vital role and can relieve staff members from performing this process manually.

Roboyo delivered multiple robots using the UiPath Hyperautomation suite to validate data inside hundreds of files with thousands of entries on Oracle E-Business Suite and Oracle Fusion Apps. Additionally, the robots reported and logged every task in a file that was stored on SharePoint, making it possible to complete time-intensive processes within hours compared to years of hours needed if the processes were to be run manually.

The first automation Roboyo developed for the company was targeted at automating the EP report validation workflow. On a high-level, the robots performed four main steps. First, they extracted ledger data from reports, then calculated and determined the difference between specific columns. The automation would generate individual worksheets for variances and discrepancies found per ledger, and last, the robots uploaded the final report to a synced SharePoint directory on OneDrive.

Another automation developed for the company’s Oracle Cloud migration project was general ledger (GL) balance spot-checking. The workflow involved 3 different robots: Dispatcher Robot, Performer Robot & Reporter Robot. The Dispatcher robot is tasked to read the input spreadsheet and push each row to an Orchestrator job queue meant to spot check GL balances. The Performer Robot will come in and review each transaction item added to the queue and start performing queries based on the account information in both EBS and Fusion.

Next, the totals for these accounts will be tallied and variances calculated. The data from the tallied results are loaded to a new job queue in the Orchestrator meant to spot check GL balance results. Data will be picked up by the Reporter Robot which will aggregate all the data, write a summary spreadsheet, and add the raw table data to an additional sheet. Lastly, the robot would save this final workbook to SharePoint.

The EP report validation automation was processing around 150 files at one time, which would only take robots around 90 minutes to complete, depending on the number of entries in each file, the highest number of entries in a file had upwards of 50,000 entries. If manually processing one single file with 50,000 rows, it would have taken more than 2,500 hours or 105 days to complete, scale that number to 150 files and the effort would have taken years to complete manually.

The automation for the general ledger (GL) balance spot-checking was also a success. The input file size would vary in size depending on what the company needed to validate. On some occasions, there were 10 to 15 accounts inside the file, but other times the robot would find more than 100 accounts. Each of those accounts had multiple ledgers that would need to be queried meaning the amount of data would grow too fast for a human to handle quickly. In contrast, the automation would be able to query more than 100 ledgers in around 40 minutes.

The largest social media company in the world is no stranger to RPA. They have created a robust Center of Excellence (CoE) to govern all automations in production and implementing strategies to source automation ideas from their employees and process experts.

The processes mentioned in this customer story are only the beginning of the client’s ambitious goals for leveraging RPA to automate the tedious tasks within the Oracle data migration and validation project.