Seal Skills Builder: Improving Contracting Party Extractions

According to a Marketscreener article, artificial intelligence (AI) is widely misused as a label for companies and products that don’t actually deploy any AI or machine learning technologies. Here at Seal Software, we routinely apply AI approaches to extract legal terms from large volumes of unstructured contract documents. With machine learning and NLP at our core, we have pioneered the Contract Analytics market and our Learning Services team helps you tailor these technologies to fit your specific business goals and contract language.

As you begin or grow with Seal,   “How to” questions frequently arise. We addressed one typical question in a previous post. Another question we are commonly asked in training classes is, “How do I get more and more relevant contracting party information extracted from my contracts?” In today’s Skills Builder, we will dive into how Seal users can easily extract this information to more effectively utilize the platform and meet their objectives.

Skills Builder Tip

If you know and can create a list of entities (contracting parties) for your contracts, you can teach Seal to recognize those names more quickly and effectively.  Seal already knows how to recognize typical references to contracting parties, but specific lists can improve recognition of less typical or non-standard language within your contracts.

To coach Seal about contracting parties, standard text files can be created containing:

  • A case-insensitive list of all company names used or contracted with,
  • A case-aware list of all company names used or contracted with,
  • A list of all personnel or entities within your company,
  • A list of all internal company names used.

These lists can often be drawn from data in an existing spreadsheet, CLM, CRM, or Accounting system.

For example:

DEUTSCHE ASSET MANAGEMENT GROUP may not wish to see DEUTSCHE ASSET MANAGEMENT (UK) LIMITED in the ContractingParties category, but instead, want this entity to be reported as an InternalContractingParty.  This can be achieved by adding DEUTSCHE ASSET MANAGEMENT GROUP to the list of “all company names used or contracted with” and DEUTSCHE ASSET MANAGEMENT (UK) LIMITED to the list of “all internal company names”.

Textual analysis enables the platform to continuously learn from the data it is given, allowing users to extract relevant information. In the example above, you are, in effect, creating subgroups to teach the system to extract the appropriate contracting party fields and you are maximizing your time to value.

See more information about creating and using these files in the “Customer-Provided Information Files” section of the Seal Contract Discovery Administration Guide. As with many of our product features, we have made some enhancements to this functionality that are further documented in the “Changes to ContractingParties/InternalContractingParties” section of the Seal Discovery & Analytics Release Notes v6.1.5 document.  We also include coverage of these lists in both the Seal Technical Specialist and Seal IT Administration courses available from Learning Services.

If you have questions you’d like answered, please send your questions and any Skills Builder Blog topic suggestions to For more information on the Seal platform and how Seal Learning Services can assist you in the skill acquisition process, visit