Thanks for reaching here! Could you share more details, on the structure of the data. And the use-case.
One thing you could possibly do is fuzzy search and define default distance as 1 or 2 to find out similar matches. But it can get tricky beyond a certain distance. Plus the whole operation would make indexing pretty slow, because you will check all the records before inserting.
Check this document on Fuzzy search - Azure Cognitive Search | Microsoft Learn
Doing the exact phone number might be easier, if you make that as the ID for the index. Otherwise Fuzzy search can help.
Let us know.