Grab is using AI agents to automate parts of its analytics workflow, reducing the share of routine analyst tickets from 44% in February to 30% in June, according to InfoQ.
The company describes mechanical work as tasks such as data preparation, alerting, and reporting. Its approach uses a five-level autonomy model to define how much of the workflow an agent can own while preserving human oversight. At Level 3, humans frame the question and review the result, while agents discover data, write and execute queries, validate results, and draft analysis.
The practical benefit is faster self-service analytics for metric, data, and SQL questions that previously consumed analyst time. The limit is equally important: Grab is not describing analysts disappearing from the process. The model depends on certified data, context management, and human review. That makes the deployment a useful example of AI changing knowledge-work queues by absorbing repetitive steps before it replaces expert judgment.