15 Sept 2019

AI Hub overview

The AI Hub is a powerful tool that simplifies the use of multiple AI engines and integrates them seamlessly into the processes built in the HighQ platform, all controlled and configured in Collaborate. The AI Hub sits between Collaborate and leading AI engines, including HighQ's own AI.

The native HighQ AI engine has been developed for document classification by type, or by language, or to determine information such as which currencies or jurisdictions are referred to in the document. You can also train the HighQ AI engine to classify new document types as well as train folders.

Access to the AI Hub includes access to the HighQ AI at no additional cost.

As there is no per-document cost to use the HighQ AI engine (for low volumes of documents), the AI Hub allows you to assess the HighQ AI engine. You can either use HighQ AI alongside other AI engines, or as a tool to reduce costs by identifying the type of file before sending them to another engine for additional processing.

Support for multiple AI engines

The HighQ platform integrates with leading AI engines such as Kira, Leverton or Eigen. These engines must be purchased and configured independently of HighQ before the AI Hub can connect to that specific service.

If you have your own AI service, you can add API support that allows the HighQ AI Hub to communicate with your AI engine. Requirements can be found on the HighQ developer community

These are the steps to enable AI and process files:

  • Contact your HighQ account manager to access the AI Hub premium feature
  • Contract access to the selected AI engines (not necessary for the HighQ AI)
  • Configure and authorise one or more AI engines at the system level
  • Turn on authorised AI engines for each site that will use AI
  • Assign the AI to a folder (this process varies for each AI engine)

See below for further details, such as storing the data in an iSheet.

Add additional AI services

Generic connector for AI

As of version 5.5, it is possible to add support for additional AI services via the generic connector for AI, as long as the service implements API commands used by the connector to send and retrieve data.

The developer community has information on the required APIs.

The AI Hub

When a file is uploaded to the HighQ Files module, it is sent to the AI Hub, analysed, key data points are extracted, and this data is then stored in the AI Hub. From the AI Hub, this data can then be used to enrich File Metadata iSheets (custom file metadata) or to use in the Data Visualisation dashboard panel. This metadata gives insights into the documents you are dealing with.

HighQ AI and the AI Hub

HighQ AI is one of several leading AI engines that may be accessed via the AI Hub; one or more AI engines can be used to analyse and tag files automatically for later reference.

After configuring access on your site, files added to folders with AI will be seamlessly analysed, and can be configured to automatically organise the analysed documents into folders.

When you first use HighQ AI, it can provide information and insights for a range of legal document types. You may provide your own example documents to further train the HighQ AI to improve accuracy and familiarity with your documents, including the ability to add your own custom tags and rules.

For example: A practice has received hundreds of documents related to a matter. These files can be uploaded to a matter folder in the Files module. The files are immediately sent to the AI Hub.
The AI Hub distributes the documents to the appropriate AI engine, and then directs the analysed files back to the Files module, ready for further consideration.

Additionally, as of version 5.6, the AI Hub can identify and extract numbers and dates from text fields returned by all AI engines.

For example, if an AI engine returned the text "This contract will automatically renew on June 24th 2021 and will run for a further 200 days if not terminated." then the user can choose to push the text to a text field (to see the full clause), or to a date column formatted as DD/MM/YYYY (in which case the AI Hub will extract the date 24/06/2021), or to a number field (in which case the AI Hub will pull out the number 200).

For example, this can be used alongside HighQ AI-trained clauses to identify and extract dates and numbers that are not part of the out-of-the-box extracted values.

How can the AI Hub help? 

Make the most of an investment without continuing costs in terms of time, resources and money; AI Hub links the Files module directly to the AI system; documents are sent automatically for analysis, either for reference or visualisation.

Use cases

AI engines assist document processing and classification for a wide range of matters:

  • Due diligence 

  • Document and contract review 

  • Compliance 

  • Contract management 

  • Knowledge management 

  • Deal analysis

  • Workflow benefits if you are already using AI

For example: You might choose to analyse documents for a contract review and tag dates, parties, locations, currencies and specific clauses.

Benefits

  • Competitive advantage - The ability to rapidly and seamlessly sort files reduces the time and effort required to manually examine documents. The time and resources saved may be re-invested in using the information gained from the AI to complete critical tasks.
  • Flexibility - You can authorise access to a multitude of expert AI systems. Choose the best AI engine for your needs to provide detailed insight and pull that information back into your central AI Hub. 
  • Automatic workflow - HighQ AI Hub handles the analysis, classification and storage automatically, freeing your team to handle other tasks and leaving you to make decisions.
  • Document analysis - Integrates clause-level analysis of legal documents with other processes without additional resources, training or overhead. Compare files and contract templates using familiar comparison tools or a numerical deviation score. You may view similar files in 'clusters' for quick analysis and management.
  • Adapts to your business - The AI can be trained on your specialist documents to tag and label documents to allow high-level analysis.
  • Cost-effective - Only analyse new files; you don't have to re-transfer all your files into the AI system and then pull the analysis back out.

How does the AI Hub work?

The AI Hub is ideal for automating and simplifying transactions such as due diligence, large property deals, legal project management, and others.

These capabilities are:

  • The AI Hub leverages HighQ’s native AI features to perform tasks such as automatically classifying document types, detecting languages, currencies or governing law.
  • It leverages powerful third-party AI engines to detect or extract provisions and data points from documents, the results of which are stored in the AI Hub.
  • The AI Hub integrates with several AI systems. After the AI machine learning engine completes its analysis (classifying, highlighting and extracting key concepts and content within the documents), HighQ stores the extracted data in the AI Hub.
  • Once this enriched information is in the AI Hub, it’s available for use in the iSheets or data visualisation modules.
    • Map AI data points to columns and automatically insert them into iSheets
    • Use data visualisation to summarise the data on a dashboard. For example, display financial data, a breakdown of documents by class and language, or show the document status.

Configuring and using the AI Hub

Assess your requirements first; a few configuration steps will activate the AI Hub and quickly provide benefits. 

Or optionally connect to an AI service via the generic connector for AI

To enable integration with a third-party AI, you must subscribe to a licence. Please discuss this with your HighQ Customer Account Manager. After the licence is enabled, contact your third-party AI representative for credentials that will allow AI Hub to connect to their AI. A HighQ system admin can use these credentials to enable the integration.

The developer community has information on the required APIs for the generic connector for AI.

A content admin can assign a template or send a file to an AI engine in the Files or iSheet modules.

Avoid assigning AI multiple times: When an AI engine is assigned to a file or folder, these are processed even if they have already been sent to an AI engine (this does not apply to HighQ AI but can affect other AI engines); this means duplicate files can appear in the assigned AI engine. 
To avoid duplication, use a consistent approach; either assign AI to files as they are evaluated ('assign as you go') or wait until it is clear how all the files should be analysed and then assign AI ('assign at the end').

Please note that if you have an AI service connected to AI Hub that supports files in addition to DOC, DOCX, TXT and PDF, you can add that file type to the AI Hub by selecting the AI Hub check box within the file admin section

Was this article helpful?