Use AI to match duplicate customer entries and reduce clutter

[HoneycombBread1]

The customer directory already has a rudimentary system for merging duplicate accounts, however it misses more than it catches.
Often, even customer entries with identical names, numbers, or emails are not identified as duplicates and in my directory of 25,000 entries, literally thousands of entries are unaccounted for duplicates. Not only does it cause problems with simple indexing, it causes trouble for our loyalty and marketing systems as well.

Is there a way to use AI to better identify and clean up these entries? Not only for exact matches, but near-duplicates as well; such as when a customer makes a typo when putting in their name or number, for example "Benjanim Smith" and "Benjamin Smith".
Would be a huge help!!