Duplicates records are part of enterprise life especially when it comes to CRM. Name prefix and suffix aside Robert does become both Rob and Bob. Richard was mostly Rick or Dick but then we all know individuals who even stuck to Hick. And we haven’t yet spoken about Asian names or for that matter Indian names where Indraneel could become Neil, Neel or even Indra. Add addresses in this duplicate mix with every city having a Chestnut, M.G (after Gandhi), Bahnhofstrasse depending on parts of world we live in, and we have a full-blown duplicate situation at hand.
Using SQL or rules to identify duplicates individually and then merging them do work but demands enormous effort to clean-up. Sainapse Intent Extraction is native designed to identify names and addresses (SSNs, telephone and credit card numbers are structured and lot easier) and goes a step further spot the groups. Sainapse can be set up to bring Robert Paulson, Rob Paulson, Bob Pulson (missing ‘a’), R Paalson (typo of ‘a’ replacing ‘u’) with very similar sounding addresses together and ask for validation and next step action.
Sainapse learns from such de-dup effort and before going live next time with new data set would already spot these groups and enable clean-up that has eluded most CRM installations.
Sainapse achieves de-dup by bucketing unstructured information via operating in ‘space of embedding’ strengthened by potential functions. De-dup set up of Sainapse can be especially effective for large customer base situations as well as powering up citizen support across City, State and Federal use cases.