November 2020 Release Notes

Number of Items:
83
Category :
Enhancement
Number of Items:
48
ID Summary Resolution
2393 Updated top toolbar with Setup, Preferences, and My Profile access

The top toolbar has been modernized and reorganized to make key features accessible while adding more space for tables, assignments, and searches in the left pane. The Setup menu has been moved to a new gear icon on the top toolbar, and a new drop-down user menu holds Preferences and My Profile along with the Language selection menu and the Logout button.  




You can adjust the top toolbar settings by going to Setup  > System > Customize Header Menu. There, you can choose whether to show the KB name, Help icon, or user profile image. You can also specify the URLs for the primary Help resource, KB-specific documentation, and more. Users can also edit their individual preferences.

2392 Pop-ups replaced with modals
Pop-up windows have been replaced with modern modal dialogs to make them easier to use, help keep the user organized during nested tasks, and prevent window saturation on the taskbar. Modals appear as an overlay within the same browser window, where they can still be resized and moved around as needed—no more stacks of windows to navigate between. The header includes the full breadcrumb path taken to reach the current editing window, so you can easily reference the other edits you're working on, and you can click any previous step in the path to jump back to that point.

If you resize or move a modal, that size and position is saved for the next time you open one. Modals also work better on mobile devices than pop-up windows, and they allow better use of the system in browsers without pop-ups enabled.



To modify the appearance of the new pop-ups, go to Setup > Look and Feel > Power User Interface, edit a scheme, and go to the Body tab. You can adjust the title of the pop-up, the current active item in the breadcrumb path, and the past inactive items in the breadcrumb path.

During long or ongoing processes, such as progress windows while Mass Edit is applied, you can use the pop-out icon to open a new window while the process completes. This allows you to keep working in your KB while long processes run. A few pop-up windows, like Formula Help, have not been converted to modal dialogs yet.


2401 Make next steps more obvious with primary and secondary action buttons
Draw users' eyes to the right action buttons by designating buttons as primary or secondary. You can use this option to make the most common buttons more obvious, and guide users more intuitively through your workflows.



By default, all action buttons are marked Primary, and both primary and secondary action buttons have the same appearance. To take advantage of this new feature, you need to designate some action buttons as secondary, and update the Look and Feel:

  1. Edit the action button you want to show as secondary.
  2. Choose Secondary as the Button Type in the "This field is displayed as" section.
  3. Go to Setup > Look and Feel.
  4. Select either End User Schemes or Power User Interface.
  5. Navigate to the Forms tab.
  6. Scroll down to the Form Primary Action Buttons section. This and the following Action Buttons sections allow you to configure button appearance.
When you edit the Look and Feel, you can also configure the appearance of primary and secondary navigation buttons, such as Save, Cancel, Next, and Finish. Navigation buttons are coded as primary or secondary, so you can't change a navigation button from one to the other, but changing their appearance to match the action buttons can be helpful to users.
2308 New look for warning and error messages
Warning and error messages have a new look, with clearer, eye-catching boxes and icons.

The appearance of these messages can also now be adjusted individually, with a new Error Messages and Warnings sections on the Forms tab of the Look and Feel editor.


2383 Input instructions shown in pop-ups to keep layouts tidy
Input instructions no longer take up extra space or appear out of alignment on record forms. Instead, input instructions are now shown in pop-ups, either over the field label or over a question mark icon next to the label. If you choose to show the pop-up over the field label, the label is underlined to indicate that it has input instructions.

The pop-up location is determined by the Look and Feel scheme of the user. To set the pop-up location, go to Setup > Look and Feel, select the interface, and edit a scheme. Then, on the Forms tab, go to the Pop-up Hints section and select the location in the "Show Hints On" field.

If you previously used code in the input instructions in order to insert an icon with a tooltip, you need to update them to contain only the instruction text. Otherwise, your existing icon will be nested inside the native pop-up, which might not work properly or might be confusing to users.
 

2376 About Agiloft page updated to reflect new release numbering scheme
The About Agiloft page has been updated to reflect the new release numbering convention, and begins with release 21. This follows release 2019_02.


2390 Introducing the Agiloft Add-in for Microsoft Word The new Agiloft add-in for Microsoft Word allows you to integrate an Agiloft KB with Microsoft Word, meaning you can import and export data between your KB and a Word document directly. This add-in was created to draft, import, negotiate, and analyze text found within both new and existing contract documents using Agiloft. Once deployed, the new add-in appears in Word on the right side of the ribbon in the form of an orange gear. You can use the Agiloft add-in with or without AI enabled.



The add-in is primarily used for analyzing Contract documents. When you upload a Contract document to Word, you can use the new add-in to extract clauses and metadata from the document if you have AI enabled. You can also manually identify clauses and metadata easily yourself, which doesn't require AI. You can use the add-in to categorize and vet the metadata and clauses included in the contract, and then send it directly to your Agiloft KB, where the information gets stored as records. You can also use the add-in to directly compare and add vetted clauses and metadata from your KB to the Contract document. For more information about using the add-in, see Using the Word Add-in

Before you can begin using the Word add-in, you must properly configure it with your KB. This ensures that the add-in knows the correct tables and fields for importing and extracting data. You can configure the add-in by visiting Setup > Integration and then clicking Configure under Word Add-in. For more information on configuring the add-in, see Configuring the Word Add-in.
2415 Automatically update Contract record values from attached file edits
Until now, if contract details were changed during the approval or review process, someone had to update those details in the Contract record to keep them in sync with the document. Now, if you use a print template to generate the document, or if you use the Agiloft add-in for Microsoft Word to manually tag metadata elements, the system can automatically keep contract metadata and clauses up to date. For example, if the Expiration Date is updated in the document, that new Expiration date value can be pushed to the Expiration Date field in the Contract record that represents the contract document.

The Print Template wizard has new options that allow you to configure print templates to push updates back to preexisting records in Agiloft. Select the "field and formula references" checkbox to track metadata updates, or select the "tag references" checkbox to track clause updates. These new options assign content controls to clauses and metadata in a Microsoft Word file based on the field and formula references within the text of the document, so you can push those changes back to the record when you finish working with the file. Additionally, if a document is created this way and then opened using the Word add-in, the metadata and clauses tagged by the Print Template wizard automatically show up on the Annotations page of the add-in.

We recommend additional configuration for this feature to add redlining for clause updates using a script action. With this configuration in place, you can push the data back to Agiloft in two ways:
  • Use the "Sync to Agiloft" button in the Word add-in.
  • Save the file and use an action button in the Attachment record to update the metadata. The clause data is updated automatically when the file is saved, using a rule.



2436 Easily request access to published AI models
Easily request access to published AI models right from the AI setup page instead of requesting access through the Support portal. Previously, users needed to request access via Support, and then wait for their Access Key and Secret Key. Now, an admin can instantly request access to generic models and obtain the required Access Key and Secret Key by clicking the "Get Keys" button. Creating Machine Learning actions that work with the generic NER, CE and Classification models is only possible after the Access Key and Secret Key have been added to the AI Credentials record. 

You can access the AI Credentials table and the Get Keys button by following Setup > Integration > AI.



2367 Automatically deploy or reject new trained AI models
The AI Model table in Agiloft now contains a new rule that, when enabled, no longer requires users to select Deploy Trained Model or Reject Trained Model after an AI Model has a status of Complete following a training session. Instead, this rule automatically deploys or rejects the model based on predetermined success metrics. For example, if a new AI Model's training metrics exceed the successful metric value and the metric direction is Maximize, the new model is deployed. 

For more information about the rule, the success metrics, and how to configure them, see  Training AI Models.


2382 Select specific labels to use with ML models
A new tab in the Machine Learning action wizard allows users to select a subset of the model's available labels. These selections tell the model which labels to identify and which labels to ignore when they are used to extract data from a contract document.

Previously, when a ML model was used in an action, every label would be identified by default. Being able to choose a subset of labels accelerates document processing, as the model is only working to return data that is actually needed. The Labels tab contains a list of the model's labels, and they are all selected by default. There are labels that work well with almost any contract type, as well as labels that are specifically designed for certain contract types. 

The Labels tab also has a search field for quick filtering. If your model contains submodels, the Labels tab shows a list of every submodel's label, with the submodel defined under the Group column. 




2321 Integrate Agiloft with LightTag for machine learning projects
Integrate your machine learning projects in Agiloft directly with LightTag, a tool used to optimize the document annotation component of the AI training workflow. LightTag annotation allows users to label specific clauses or metadata within a document, such as a termination clause or introductory greeting, based on text content. The labeled content is then used to train machine learning models to recognize similar labels in other documents. 
 

By directly integrating LightTag and Agiloft, you can send information, such as project data, annotators, labels, and document sets, between Agiloft and LightTag with the click of a button. These buttons run Annotation actions that are specially designed to synchronize information between the two platforms. For more information about how to configure Annotation actions, see LightTag Integration.

 

LightTag integration is useful only with an Enterprise Extended license. Without this license, the system cannot train machine learning models, which is the principal benefit of LightTag integration. To enable integration with LightTag, input relevant data in the tabs of the LightTag wizard by following Setup > Integrations > LightTag Annotations in Agiloft. The information found in the tabs of the LightTag wizard are used to configure an Annotation action, which are used to share or synchronize relevant data directly between Agiloft and LightTag.
2414 Package multiple models in one container
This release made several behind-the-scenes changes that are helpful for administrators who work with ML models. The Agiloft AI Core model container is now able to use and support ML models from Huggingface. This opens access to several hundred models, which have all been pre-trained on various different language tasks.
 
The AI Core model container can now also package multiple ML models together. One model may combine up to 20 submodels, depending on the number of labels. From the user perspective, multiple models packaged into one container appear as one large model in the KB. This mechanism has been added to solve two problems:
  • Models are now much easier to extend and maintain. Now, models can be added each time a new label is created. Existing models can also be swapped out for better models if necessary.
  • Fine-tuning model labels is also much easier. Now, models compile many different datasets each for one small number of labels, whereas before one dataset contained training data for all labels. You can use the Labels tab of the Machine Learning action wizard to determine which labels should be identified by the model.
 
 
2372 New S-BERT model for similarity scoring

Agiloft now offers a SEMANTIC-TEXTUAL-SIMILARITY model, referred to as S-BERT. This AI model evaluates semantic textual similarity of an input field against a field that contains existing data, both of which can be configured by the user.


The existing data field can either be a Text field or a linked field with multiple values, meaning you can compare the input against one value or many. For example, if you upload a contract clause to the input field, the model can compare that clause against a single clause, or against all of the clauses of the same type in the Clause Library. The model assigns potential matches with a similarity score between 0 and 1. You can choose to run the model with either cosine (default) or chebyshev metrics. For more information, visit Similarity Scoring with AI.

2407 Support ONNX +FP16+I8 inference and model conversions for NER-G2 container The NER-G2 model container has been improved by adding ONNX accelerator support and static optimizations, as well as FP16/I8 quantization for the hardware that supports it. As a result, model performance and speed for both inference and training is significantly improved.
2404 CML KB limit change to "AI Requests Model Time allowed per model per day" The Central Model Library limit per model has been changed. Instead of using the number of requests per day as the limit, the time spent using a model is now the primary limit. 

A new "AI Requests Model Time allowed per model per day" field has been added to designate that time limit with a default of 1200 seconds. It can be customized for individual users. When a new inference request arrives, it adds up all Model Response Time(s) of this client's model over the last 24 hours.
2386 ML actions can now map extracted data to File fields
You can now map data from extraction models to File fields, in addition to the previously available Text fields.

 
2400 Integration with Tableau

Connect Agiloft data to Tableau by using any saved search as a Web Data Connector (WDC). This allows you to pull the saved search data into Tableau for analysis. You can configure multiple connectors using any number of saved searches.


To set up a saved search to use with Tableau, follow the setup steps in Tableau Integration.


2395 Connect KBs to Boomi with new Agiloft connector
The Agiloft connector makes it easy to exchange data between Agiloft and applications on the Dell Boomi platform. When you've connected the two systems, you can configure your connector to do any or all of the following:
  • Pull data using Get, GetChoiceListId, Query, and Retrieve Attachment
  • Add or update data using Create, Update, Delete, Attach, and Remove Attachment

If multiple users need to work with the connector on the Boomi platform, or if one user needs to connect to multiple KBs, configure a connector for each user-KB combination. Access to KB data is controlled by the permissions attached to the Agiloft credentials you use to configure the Agiloft connector in Boomi. The Agiloft connector integrates with Agiloft using REST API calls with the JSON request/response format over an HTTPS connection.


For more information and details on setting up your KB to work with Boomi, see Boomi Integration.

2413 Real-time, record-specific Salesforce syncing You can now configure Salesforce integrations to sync individual records, in addition to the existing option to sync the entire data set. To use this record-specific option, create a Sync action with an existing Salesforce sync configuration and select the "Sync a single record only" checkbox. 

Make sure this action is used in a context where records are selected, either by selecting records in the table view and using the Actions option in the toolbar; creating an Action Button to use inside individual records; or using a rule targeting a specific record or set of records. If you use the action in a time-based rule or other non-record-specific context, no records are synced.



On the Salesforce side, you can use these URL parameters to sync a particular record: &explicittables=SF_TABLE_NAME&explicitids=SF_RECORD_ID
2410 New DocuSign authentication method for systems with 2FA
If your system is configured to use two- or multi-factor authentication (2FA) during user login, users have to authorize access with 2FA every time they use DocuSign. To avoid this, you can use the new JWT authentication configuration, which uses the Client ID, Account ID, and RSA Key listed in the DocuSign Account Admin settings. This allows DocuSign to authenticate the user instead of using the KB login credentials to do it.
 

 
To use this new authentication method:
  1. Go to Setup > Integration and click Configure under DocuSign Extension.
  2. Complete the Client ID, Account ID, and RSA Key fields with the information found in the Account Admin section in DocuSign.
  3. Click Save and navigate to the DocuSign Users table.
  4. Create or edit a DocuSign User record.
  5. In the Authenticate field, select JWT.
  6. Save the record. Repeat steps 4-6 for other user records as needed.
With this configuration, when a user accesses DocuSign, their session is authenticated with an OAuth token and they no longer need to confirm their identity using 2FA.
2399 Use REST API to trigger Actions Use a new REST endpoint, EWActionButton, to trigger an action button in the system. This calls the action button asynchronously and then requests the status of the operation. For more information, see REST - Action Button.
2381 Adobe Sign security enhancements Adobe Sign security has been improved by restricting security on two parameters passed during the OAuth flow. These parameters are now protected against changes, except from .echosign.com, .adobesign.com, and .documents.adobe.com domains.
2371 Removed Ephesoft integration Integration with Ephesoft is no longer supported, and it has been removed from the Integration setup menu.
2369 Removed Skype and Skype for Business integrations Integration with Skype and Skype for Business is no longer supported. These options have been removed from the Integration setup menu.
2397 Google OpenID 2.0 SSO deprecated
Because Google no longer supports OpenID 2.0, the Google option under Setup > Access > Single Sign-On is no longer available. Google Oauth 2.0 single sign-on is still supported and remains the preferred method for SSO with Google.
2388 Updated email and HTML editors
The style of the email and HTML editors has been given a smoother, modern update.


2391 Email signatures in Agiloft
Email signatures are now supported in outbound email accounts in Agiloft. Once signatures have been configured, they can be inserted by following Insert > Insert signature > Search... in the email editor.
 
 
To set up this new feature, first set up the KB structure to store the signatures:
Enable email signatures by following these steps:
  1. Go to Setup > Email and SMS.
  2. Click All Communications Table Outbound Behavior.
  3. At the bottom, select Email Signatures.
  4. Under Name Field, select the Signature Name field.
  5. Under Content Field, select the Signature Content field.
Administrators and users can create signatures to use in the system:
  1. Open the user menu and click My Profile.
  2. Go to the Emails tab.
  3. In the Signatures section, click New to create a new Signature record.
  4. Enter a name, set the visibility, and enter the content of the signature.
2387 New emails now search recipients by Full Name by default
When you create a new email and click the magnifying glass to look up a recipient, the default search option is now Full Name to make searching easier and more intuitive. Previously, the default search option was Login.


2385 Clearer SMTP server setup for TLS and SSL
The email server configuration now offers clearer guidance in the SMTP Server Info section by using radio buttons to indicate the type of connection. Previously, these options were all listed as checkboxes, even though only one option should be selected.


2370 Identify records with lots of History entries
Administrators now have access to a new report that shows records with a large number of History entries, which can be useful in troubleshooting space issues or problems with excessive or duplicate History entries. The report shows the total size of History data in each table, and the individual records with the most History entries.

To access the report, go to Setup > System > History Report.