Another option that might fit your use case a little better would be to leverage a serverless function, like an AWS Lambda, Google Function, or Azure Function, that is tied into their respective database function. We have a data action integration to Lambdas, and Google or Azure functions can be invoked using the Web Service Data Action integration. Each of those platforms have a concept of "Time to Live" (TTL) in their database technologies; this will allow the entry to "age out" of the data table automatically, so you don't need to have some form of a cleanup step after the successful call.
If I were designing the flow, I would do the check for the number in the Data Table, and if it's found then transfer transfer to another task that increases the priority of the call (or other actions you're wanting to do. If the number isn't found in the data table, perform another data action that places the number into the data table (in case the call abandons and the person calls back in later), then proceed with your normal call flow operations.
------------------------------
Richard Schott
Genesys - Employees
------------------------------