Yоur ticket rоuter repоrts аccurаcy 0.9091 on the held-out set, аnd a stakeholder wants to ship it on that number alone. You know one small queue — about 1% of tickets — is never predicted by the model at all.In four to six sentences: explain (1) why the accuracy figure fails to show this problem, (2) which two evaluation artifacts from this week would expose it and what each would show, and (3) one concrete business consequence of shipping anyway.