Resolution rate is calculated by dividing the number of customer contacts resolved in a single interaction by the total number of contacts handled, over a given period. A contact counts as resolved only when the customer's actual need was met, not simply when a conversation ended, so businesses typically also track repeat-contact rates to catch cases where an issue was closed prematurely but the customer came back with the same problem shortly afterward.
Tracking resolution rate separately for AI and human agents lets a business see exactly how much of its Arabic and English customer volume the AI concierge is genuinely resolving end to end, rather than just deflecting or routing. In regulated industries like banking and healthcare, where an unresolved issue can mean a missed payment deadline or a delayed prescription, this distinction between contacts closed and needs actually met is a core measure of whether AI automation is working.
In Eshal: Eshal reports resolution rate separately for its AI concierge and any human-assisted contacts, giving customers visibility into how much volume the AI is resolving end to end versus escalating. Eshal's average automated resolution rate across deployments is 86%, tracked alongside overall resolution rate to distinguish contacts the AI fully closed from those it handled without full resolution.