{"slug":"freshworks-uses-machine-learning-to-predictively-triage-help-desk-tickets","url":"https://findausecase.com/use-cases/freshworks-uses-machine-learning-to-predictively-triage-help-desk-tickets","title":"Freshworks uses machine learning to predictively triage help desk tickets","description":"Freshworks, a CRM and customer experience software provider with over 60,000 enterprise customers, migrated from Hadoop/Cloudera to the Databricks Data + AI Platform to use machine learning for predictively and automatically triaging help desk tickets, greatly increasing ticket resolution speed. Using MLflow, its data science team streamlined the ML lifecycle from training to deployment, training models 4x faster after the migration, while reducing data platform maintenance costs by 75%.","company":"Freshworks","industry":"Technology & Software","aiCapabilities":["Predictive Analytics"],"technology":["Databricks SQL","Delta Lake","MLflow","Unity Catalog"],"deployment":"Public Cloud","problemStatement":"Freshworks' legacy Hadoop infrastructure (self-managed CDH distribution of Cloudera), made up of multiple data and analytics tools, incurred massive IT overhead and created performance bottlenecks as data volumes increased, slowing the customer support team's ability to triage and resolve help desk tickets.","solutionApproach":"Freshworks migrated to the Databricks Data + AI Platform over seven months, using Delta Lake for data reliability, Unity Catalog to democratize data while meeting PII compliance requirements, Databricks SQL for operational analytics, and MLflow to streamline the machine learning lifecycle so it could predictively and automatically triage help desk tickets.","businessValue":"Freshworks reduced data platform maintenance costs by 75%, boosted data team productivity over 60%, achieved 4-5x faster data processing, and trained ML models 4x faster, leading to faster help desk ticket triage and higher customer satisfaction.","evidence":{"band":"high"},"sourceUrl":"https://www.databricks.com/customers/freshworks","dates":{"publishedAt":"2026-09-25T05:47:43.814Z","publishedAtSource":"pipeline","updatedAt":"2026-09-25T05:47:43.814Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/freshworks-uses-machine-learning-to-predictively-triage-help-desk-tickets. Bulk republication requires permission."}