Explore LABS


Where's the AI?

Previous Article     Index        Next Article 

We’ve imitated the natural intelligence humans use when comparing contact information.


To do this, first we need to do sift through the data to find the records that a person would want to review – without needing a data scientist to pre-process and standardise the data. We appy Natural Language Processing (NLP) methods to the typical challenges presented by contact data. Applying NLP techniques like lexical semantics to people and business names, addresses etc. matchit develops an understanding of your data based on what it is and not where it resides in a table.

For example, matchit understands that Tony is a first name, short for Anthony; Mr. is a title or prefix; MIEE is a qualification; PAS is an acronym of  Palmer Alarm Systems; Operations Director is a job title despite being in the wrong column; Kings Mill Ln is a thoroughfare and Redhill is a town; RH1 5YP is a postcode in the wrong column; @pac.co.uk represents a corporate email whereas @gmail.com will be a personal email address. It even knows that info@ will be a departmental email address and x207 is a telephone extension number.

Not only does matchit match poorly structured data easily, it’s amazingly tolerant to wide variations in customer data. Its Smart Matching logic uses all the contact information available to find matches by default, so if there are significant items of data available such as email, phone, date of birth, name, or address, it uses them automatically. If this data isn’t present in some records, or different emails and phone numbers were used at various times – no problem. The bottom line is that matchit figures out the best matching rules for your data so you don’t have to be an expert!

Our Smart Matching technology is available within the following products:

  • Cortex
  • matchIT Hub

and you can find more information in our knowledgebase by clicking through to the following location - Smart Settings in detail.

Previous Article     Index        Next Article 


Was this article helpful?
0 out of 0 found this helpful


Please sign in to leave a comment.