For enterprises managing millions of customer interactions, ensuring compliance is a daunting yet essential task. Flutter, a major player in the entertainment industry, faced this challenge within its web chat customer service. With over 5 million conversations per year, manually reviewing each one for compliance with security and verification procedures was inefficient and resource-heavy. IBM’s Client Engineering team stepped in to co-create a smarter, automated approach using AI and automation technologies.

By leveraging Cloud Pak for Data, Watson NLU, and Cloud Pak for Business Automation, IBM developed a six-stage Python data pipeline that revolutionised Flutter’s compliance process. The pipeline automatically:

• Extracts conversation data via API from LivePerson
• Cleans the data using spellcheck and text processing
• Performs sentiment analysis with Watson NLU
• Identifies compliance issues using custom AI models
• Prioritises quality assurance cases with IBM Operational Decision Manager (ODM)
• Feeds insights into a Cognos-powered dashboard for easy review

With this solution, compliance teams can now dynamically identify and prioritise agents for review, reducing manual workload and improving oversight. The Cognos dashboard provides a structured, visual representation of data, enabling teams to track compliance trends and quickly spot potential issues.

The results speak for themselves: a fully automated, AI-driven compliance system built in just five weeks, drastically improving efficiency and ensuring security procedures are consistently followed.