Network management systems expose two interfaces. A web UI for operators and a REST API for integrations and scripting. They're supposed to behave the same way. In practice, they often don't. A VLAN configuration that works through the dashboard might fail silently through the API. An interface status that shows correctly in the API might render wrong in the UI. Most automation projects only cover one surface, so the other one goes untested and those inconsistencies make it to production.
We build unified automation suites that exercise both surfaces against the same scenarios. The same test that provisions a device through the web UI also provisions it through the API, and verifies that both paths produce the same result. We use Selenium, Playwright, and direct REST/SNMP/NETCONF calls, whatever your NMS exposes. The suites run in CI, and the results show a clear picture of where the two interfaces agree and where they diverge.
If there's a mismatch between what your customer sees in the dashboard and what their script gets from the API, you want to catch that in the lab. Not in a support ticket.