Automated resolution rate is calculated by dividing the number of contacts fully closed by the AI, with no handoff to a human agent, by the total number of contacts the AI handled. It is typically tracked per channel and per use case, since resolution is easier for simple, well-defined requests like order status than for complex or emotionally sensitive issues. Systems that can only answer questions tend to sit at the lower end of the range, while systems that can execute actions in connected systems reach the higher end.
In MENA deployments spanning banking, healthcare, retail, logistics, and government, automated resolution rate varies significantly by vertical and query complexity - a logistics tracking request is far simpler to fully resolve than a complex healthcare booking. Because the metric is measured consistently across deployments, it lets businesses benchmark AI performance against industry data such as the AI Customer Experience Benchmark, and set realistic expectations for what a WhatsApp-based AI concierge can close without escalation.
In Eshal: Eshal's average automated resolution rate across deployed customers is 86%, ranging from 78% for complex healthcare bookings to 93% for logistics tracking, consistent with the AI Customer Experience Benchmark 2026. This range reflects that some workflows, such as multi-step KYC or clinical scheduling, inherently require more verification steps than a straightforward shipment status check, even when the same underlying agent handles both.