In October 2020, NBN engaged Parplex, a technology firm that specializes in robotic process automation (RPA) software solutions. Ellen Bartholomeeusen, NBN’s Operations Manager, says, “It was important for us to see that Parplex was an IBM Business Partner. And we choose Parplex because they proposed the best solution.”
The collaboration began with a pilot focused on processing a small subset of ballots. In January 2021, after the pilot’s success, Parplex turned its attention to automating the entire process of implementing ISO standards as Belgian standards and configuring the bot, called Otis. Parplex worked iteratively and fine-tuned and validated the bot for each phase. After end-to-end testing, NBN launched Otis in May 2021.
Nordine Vandezande, the owner of Parplex, elaborates: “IBM RPA software has all the features necessary to build a bot and automate NBN’s ballot process. We configured the complete process: creating the ballots and collecting the information from the standards. The bot runs every evening. It functions based on spreadsheets that contain ISO-related metadata, such as the number and description of the standard. By reading the spreadsheets, the bot knows which ballots it needs to create and which ones it needs to close.”
Otis runs on a workstation completely configured by Parplex. It integrates with the in-house product information management (PIM) system, the ISO website and NBN’s software-as-a-service (SaaS) platforms. This ensures that the information gets where it belongs in real time.
Ms. Michel adds: “Each morning we immediately see which ballots were implemented correctly — they’re flagged with ‘ok.’ In some cases, the bot can’t fill in a field because the information is missing in the ISO systems or for some other reason. In these cases, the bot highlights that field and human intervention is needed to make a correction or fill it in.”
Ms. Bartholomeeusen also appreciates the solution’s ability to report errors and bugs. “It makes the task easier by showing us what did not work so we can finish the task very quickly,” she says. “I’d say that only 5% of the standards implementations require manual intervention. From a speed and quality point of view, implementing the bot has been a great added value.”