AI agents gain autonomy faster than enterprise verification

Half of enterprises report AI agent failures after passing internal tests, with one in four experiencing multiple such incidents. This evaluation gap undermines trust as agents gain more autonomy faster than companies can verify them.