{"id":20644,"date":"2025-09-18T18:08:26","date_gmt":"2025-09-18T18:08:26","guid":{"rendered":"https:\/\/www.directimpactsolutions.com\/?p=20644"},"modified":"2025-10-08T13:04:24","modified_gmt":"2025-10-08T13:04:24","slug":"filemaker-ai-and-entity-matching-pt-2","status":"publish","type":"post","link":"https:\/\/www.directimpactsolutions.com\/en\/filemaker-ai-and-entity-matching-pt-2\/","title":{"rendered":"FileMaker AI and Entity Matching \u2013 Pt. 2"},"content":{"rendered":"

A technical approach for leveraging Semantic Find to solve the difficult problem of fuzzy matching<\/h4>
<\/path><\/svg><\/div>

UPCOMING EVENT: Using AI to Handle Duplicates in FileMaker<\/strong> <\/h3><\/div>

Join Direct Impact Solutions for an exclusive live webinar to learn how AI can simplify searching your data, detect and clean duplicates, and prevent them before they happen.

\ud83d\udcc5 Date:<\/strong> September 30, 2025
\u23f0 Time:<\/strong> 11:00 AM Pacific \/ 2:00 PM Eastern \/ 8:00 PM Central Europe<\/strong> 
\ud83d\udd17 Reserve your place here:<\/strong> Webinar Registration<\/a>

Host: <\/strong>David Weiner, Senior Developer and Project Manager at Direct Impact Solutions
Languages:<\/strong> English audio with simultaneous French subtitles

Don\u2019t miss this unique opportunity to interact with our experts and explore the full potential of AI in FileMaker for your organization!<\/p><\/div><\/div>

\"\"<\/figure>

In part 1 of this article<\/a>, I described how FileMaker\u2019s Semantic Find<\/strong> differs from its regular find, what Entity Matching<\/a> is (and why it\u2019s important for FileMaker users), and how Semantic Find can make deduplication tasks easier. Here in part 2, I\u2019ll cover the specifics of an entity matching approach in FileMaker using Semantic Find to perform fuzzy matching. I\u2019ll only briefly touch on how to set up Semantic Find in a FileMaker solution, since there are plenty<\/a> of other<\/a> articles<\/a> covering this topic in detail, but I will<\/em> go over the design choices specific to entity matching.<\/p>

What Is Entity Matching?<\/h2>

There\u2019s a great overview of Entity Matching at the link above, but if you aren\u2019t familiar with it (and you haven\u2019t read part 1 of this article), it\u2019s the process of determining if two records refer to the same real-world entity. It\u2019s required in a wide variety of contexts, and is crucial for the deduplication of data. It has also been a continual and vexing problem in the FileMaker systems I work with daily.<\/p>

The basic problem is to be able to compare two (or more) records and figure out, programmatically, whether they\u2019re actually referring to the same thing. Is \u201cHanna C Beuler\u201d the same person as \u201cHannah Bueler\u201d?<\/p>

Record 1: Hanna C Bueler, born on 03\/12\/1990<\/strong><\/p>

Record 2: Hannah Beuler, born on 12\/03\/1990<\/strong><\/p>

In FileMaker, using the normal \u201cexact match\u201d find, it would be difficult to even locate<\/em> these two records to compare to each other (since their spelling and dates of birth are different), let alone determine if they\u2019re a match. A human user could see pretty quickly that these are likely the same person, but a simple FileMaker search can\u2019t do that. This is where Semantic Find comes in.<\/p>

Looking for a Match<\/h2>

The usual process of looking for duplicates would be to search multiple fields (or some kind of concatenated calculation field) for an exact match:<\/p>

Enter Find Mode [ Pause: Off ]\nSet Field [ Person::FirstName ; \u201cHanna\u201d ]\nSet Field [ Person::LastName ; \u201cBueler\u201d ]\nSet Field [ Person::DOB ; \u201c3\/12\/1990\u201d ]\nPerform Find [ ]<\/code><\/pre>
Perhaps this goes without saying, but if you find what you\u2019re looking for using this type of standard search, then you don\u2019t need to do anything with semantic find! You\u2019ve found a match, and you can move on. However, if your initial search turns up no exact matches, it\u2019s time to try semantic find to locate a close (but not exact) match. Semantic find is a game-changer for finding tough-to-match records once you\u2019ve exhausted the simple \u201cexact match\u201d find.<\/p>
Setting up FileMaker for Entity Matching<\/h2>
In order to perform entity matching using Semantic Find in FileMaker, we first need to create vector embeddings<\/strong> for our match data.<\/p>
Embeddings<\/h2>
\"\"<\/figure>
A vector embedding is a numerical representation of data that captures its semantic meaning. We store embeddings of our search data in a container field. FileMaker\u2019s Semantic Find then compares these vectors with your search query, returning a similarity score. So, the first thing to do is decide what data you\u2019d like to match against and add container fields to store their embeddings.<\/p>
While embeddings can be stored in the search table itself, I prefer to store them in a separate table in a one-to-one relationship with their data table.  Using FileMaker 22\u2019s built-in support for either OpenAI, Anthropic, or Cohere, this is as easy as using the \u2018Insert Embedding in Found Set\u2019 script step and choosing the fields you\u2019d like to create embeddings for:<\/p>
Insert Embedding in Found Set [ Account Name: \u201cMy_Semantic_Find\u201d ; Embedding Model: $$loaded_model ; Source Field: Person::ConcatenatedFirstLastDOB ; Target Field: EmbeddingTable::VectorizedFirstLastDOB ; Replace target contents ]<\/code><\/pre>
I opted to use Google\u2019s Gemini embedding model<\/a> instead of one of the built in model providers, since<\/p>
  1. I like Google\u2019s tools<\/li>\n\n
  2. The model\u2019s performance and accuracy are excellent<\/li>\n\n
  3. The embeddings are half the size of OpenAI\u2019s (6k vs. 12k), and<\/li>\n\n
  4. Their embedding model is free<\/a><\/li><\/ol>
    Using Gemini requires a bit more work, since it requires making a REST call with the \u2018Insert from URL\u2019 script step, parsing the resulting vector data, and storing the embedding in a container field as a binary file (using the \u2018GetEmbeddingAsFile\u2019 function), but it\u2019s not too difficult if you\u2019re familiar with making API calls in FileMaker:<\/p>
    \"\"<\/figure>
    Once we\u2019ve added our embeddings, we can begin to search against our data table using Semantic Find.<\/p>
    Embedding considerations<\/h2>
    Entity matching is different from an ordinary semantic find, since we\u2019re not so much looking for similarity in meaning<\/em> as we are looking for just plain similarity. We want to match \u201cHannah Beuler<\/strong>\u201d with \u201cHanna C Bueler<\/strong>\u201d, not with [several paragraphs about the life of Hanna C Bueler<\/em>]. This being the case, we only need to create vector embeddings of a selection of fields used for matching. And since the fields we select will typically be pretty small (as opposed to large blocks of text), the token count<\/a> will remain small as well.<\/p>
    So how do we pick what to vectorize for entity matching? There are probably any number of ways to approach this, but the method I\u2019ve used with success is to create a calculated field of a short match phrase that, if identical to your query, would represent a perfect match. So, if you\u2019re searching for a person, you might create a calculation field consisting of FirstName, LastName, DateOfBirth, and PhoneNumber. If you\u2019re searching for an organization, you might use OrganizationName, Address, and PhoneNumber. For a catalog product, you might use ItemDescription and Size. Here are some basic guidelines:<\/p>