aiR for Case Strategy Limited General Availability

This application uses powerful Generative AI to expedite the building of your case strategy. With aiR for Case Strategy, you can generate facts automatically from key documents, get customizable document summaries, and use large sets of these facts to craft witness summaries and deposition outlines to prepare for interviewing individuals.

Before you begin

Install the Case Dynamics application before using aiR for Case Strategy.

Installing aiR for Case Strategy

aiR for Case Strategy is available as a secured application from the Application Library. You must be a participant in the Limited General Availability program. aiR for Case Strategy is not available for repository workspaces.

To install the application:

1. Navigate to the Relativity Applications tab in your workspace.
2. Select Install from the application library.
3. Select the aiR for Case Strategy application.
4. Click Install.

For more information on installing applications, see Relativity applications.

Permissions

The following permissions are needed to use aiR for Case Strategy:

Object Security
aiR For Case Strategy Case Document - View Edit Add aiR for Case Strategy Prompt Criteria - View, Edit, Add Fact - View, Edit, Add

Generate facts

You can turn a set of documents of interest into a group of relevant facts containing concise document summaries. These facts can then be viewed on a chronology, gathered to prepare a witness summary or a deposition outline, or used in customizable workflows.

You can generate facts using aiR for Case Strategy either by selecting a saved search from the Case Home tab or by selecting documents on the Documents tab.

To generate facts using a saved search on the Case Home tab:

1. Navigate to the Case Home tab.
2. Click Create a Fact Chronology.

3. Select a Saved Search to specify which documents are used to generate facts.
   Note: To ensure high performance, there is a limit of 5,000 documents at a time. To learn more, see Size limitations.
   
4. Complete the fields in the modal. To learn more about the fields, see Prompt criteria below. For tips on how to write for the prompts, see General writing guidelines.
   Note: Each Generate Facts field has an individual limit of 4,096 characters except for the Prompt Criteria Name field which has a limit of 256 characters.
   Note: If you plan on using information entered in the fields again in the future, we recommend copying and saving it in a text editor or word processor to make it easy to reuse.

5. Click Generate Facts.

To generate facts by selecting documents on the Documents tab:

1. Navigate to the Documents tab.
2. Manually select the documents to use for generating facts. The mass operations bar, which allows you to select all documents in the list, is not currently compatible with aiR for Case Strategy.
   Note: To ensure high performance, there is a limit of 5,000 documents at a time. To learn more, see Size limitations.
3. Click the Generate Facts button.

4. Complete the fields in the modal. To learn more about the fields, see Prompt criteria below. For tips on how to write for the prompts, see General writing guidelines.
   Note: If you plan on using information entered in the fields again in the future, we recommend copying and saving it in a text editor or word processor to make it easy to reuse.

5. Click Generate Facts.

Prompt criteria

The prompt criteria in aiR for Case Strategy give the large language model (LLM) the broad context surrounding a matter. This includes an overview of the matter, people and entities involved, and any jargon or terms that might help in understanding the documents.

Most fields display guidance on the voice and tone of what you should write. Further description of the fields for generating facts is provided below.

• Prompt Criteria Name—Enter a name for this version of the prompt criteria. The name should be unique between jobs, and will be filterable when you are navigating the list of facts.
• Matter Overview—Enter a concise overview of the case. Include the names of the parties involved, the nature of the dispute, and other important case characteristics.
• What are you trying to prove?—Outline your side of the argument and what you are trying to achieve in the case. This helps aiR for Case Strategy judge each fact as helpful or harmful to your case.
• Issues—Enter a list of issues and their descriptions. Issues should be in the following format: Issue name: Description. The issues you enter are used by aiR for Case Strategy to tag the facts that it creates.
• People and Aliases—Enter the names and aliases of key individuals in the case. Include their role and any other affiliations.
• Noteworthy Organizations—Enter the name of organizations and other relevant entities involved in the case. Highlight any key relationships between the entities or their notable characteristics.
• Additional Context—Enter any additional information that does not fit the other fields. This section is typically left blank.
• Document Summary Instructions—Describe your desired content, structure, and/or formatting for the document summary. If you leave this blank, the job will create readable summaries of the document, but content, structure, formatting and style are not explicitly prescribed.

aiR Fact fields

Once facts are generated, a set of fields are populated with information based on your selected documents. aiR for Case Strategy creates up to 9 facts per document and the AI looks for any information that is relevant to your prompt criteria. Each fact is tied to a single document and duplication of facts can occur if they are generated from documents that are similar.

The following fields are populated after generating facts:

• Primary Fact Date—denotes the date when the fact occurred, or the start date for an event that occurred between certain dates, including the time.
• Fact—a brief title for the fact.
• Description—a description of the fact in greater detail.
• aiR Issues—displays the issues that aiR for Case Strategy has associated with each fact. The issues used are from the info you provide in the Issues field on your prompt criteria.
• aiR Entities—displays the entities that aiR for Case Strategy has associated with each fact. You do not have to mention people in the People and Aliases field in order for them to be found and set in the aiR Entities field.
• Document Summary—displays an AI-generated synopsis of the document which this fact came from.
• Fact Documents—displays the control number for the document which this fact came from. You can link a fact to more documents manually to also display multiple control numbers in this field, but document preview will only display the first document listed.
• Citation—the portion or portions of the document that is used to support each fact. If multiple excerpts were found, they will be separated by commas.
• Fact Score—an AI-generated score for each fact on a 0-4 scale.
  • 4—Very important; the fact is crucial and has a strong impact on proving the user's case.
  • 3—Important; the fact has a significant impact and is very relevant to the case.
  • 2—Moderately important; the fact has some impact and is somewhat relevant.
  • 1—Slightly important; the fact has minimal impact but might still be somewhat relevant.
  • 0—Not important; the fact has no relevance to the case. This score is rare, as aiR for Case Strategy usually only extracts facts that it deems relevant.
• Arguments for Fact Score—an AI-generated rationale for each Fact Score to help you understand the scores better.
• Helpful/Harmful—an AI-generated rationale on whether a fact is helpful, harmful, or is neutral towards your case.
• Keep Fact—a Yes/No field which you can use for QC'ing facts and recording whether a fact is useful without having to delete it.
• Prompt Criteria Name—the name entered in the Generate Facts modal. This field has a character limit of 256 characters.

Fact generation considerations

Fact generation uses Generative AI to read the extracted text of the documents, describe the contents using a document summary, and to create and score the facts. We strongly recommend that you provide an appropriate amount of oversight by having an attorney working on the case review the facts and then confirm or correct their language as needed. For further Generative AI considerations, see the Considerations regarding the use of Generative AI section below.

If the same facts occur in multiple documents, duplicate or overlapping facts can be generated. To prevent this, avoid including duplicate content in the documents you select for fact generation.

Running multiple fact generation jobs on the same document can also result in duplication, as prior facts are never deleted. We recommend that you use the Prompt Criteria Name field to filter by run or delete old facts if you have made significant improvements to your prompt criteria.

There is not currently a way to load prompt criteria from a prior run. We recommend that you use a text editor or word processor to preserve prompt criteria for future use.

Editing facts using inline editing

Users can use inline editing on the Case Home tab to edit cells on-the-fly without needing to open each fact individually. Eligible fields for inline editing are identified with a yellow highlight at the top of the column. To learn more about how inline editing works on the Documents tab, see Inline editing in the document list.

Click the Inline Editing icon to enter and exit edit mode. While in edit mode, you can edit any eligible fields.

aiR Facts view

Once you've generated facts using aiR for Case Strategy, you can view the results using the new aiR Facts view on the Facts tab. aiR Facts is a default view and cannot be edited.

Optionally, you can create a copy of the aiR Facts view and add or remove fields to create a custom view that suits your needs.

To create a custom view, do the following:

1. Navigate to the Views tab.
2. Select the aiR Facts view from the list.
3. Select Copy from the mass operations toolbar.
4. Click Ok.
   
   The aiR Facts (1) view is created and it is an editable view that contains all of the aiR Facts fields which you can further customize.
5. Select the aiR Facts (1) view from the list.
6. Click Edit.
7. Edit the Name field to ensure your new custom view has a recognizable name.
8. Ensure the Fields tab is selected and then add or remove fields as desired from the Selected section. Each field in the Selected section will display in the custom view.
9. Once you've finished editing the fields, click Save.
10. Navigate to the Facts tab.
11. Click on the Views drop-down menu and select the view you created to begin using it.

Create a deposition outline

Once you have generated a collection of facts, you can use them to build a deposition outline. This document provides a starting point for an attorney to plan a witness interview, using both the input facts, information you provide about the most important issues, and your own objectives in the case to construct an appropriate plan.

To create a deposition outline:

1. Navigate to the Case Home tab.
2. Select the facts you would like to use to create the deposition outline from the list. Facts must contain a date and a witness in its aiR Entities field to be usable.
   Note: You can select up to 400 facts when creating a deposition outline.
3. Click Create a Deposition Outline in the aiR for Case Strategy panel.

4. Complete the fields in the modal. To learn more about the fields, see Prompt criteria below. For tips on how to write for the prompts, see General writing guidelines.
5. Click Create Deposition Outline.
   Once the deposition outline is created, you can view it by selecting the deposition outline in the Case Documents widget. A deposition outline can also be edited, to learn more see Editing a deposition outline or witness summary.

Create a witness summary

Once you have generated a collection of facts, you can use them to build a witness summary. This document provides a snapshot of a witness’s involvement in a matter, using the input facts, information you provide about the most important issues, and your own objectives in the case.

To create a witness summary:

1. Navigate to the Case Home tab.
2. Select the facts from the list that you would like to use to create the witness summary.
   Note: You can select up to 400 facts when creating a witness summary.
3. Click Create a Witness Summary in the aiR for Case Strategy panel.

4. Complete the fields in the modal. To learn more about the fields, see Prompt criteria below. For tips on how to write for the prompts, see General writing guidelines.
5. Click Create a Witness Summary.
   Once the witness summary is created, you can view it by selecting the witness summary in the Case Documents widget.

Prompt criteria

Most fields display guidance on the voice and tone of what you should write. Further description of the fields for creating a deposition outline or witness summary is provided below.

• Prompt Criteria Name—Enter a name for your prompt criteria responses. This name should be unique between jobs.
• Entity—Enter the name of the person that this deposition outline or witness summary is being created for. The Entity will be matched with the Key Entities field to determine if a fact should be considered in the generation of a deposition outline or witness summary.
• Matter Overview—Enter a concise overview of the case. Include the names of the parties involved, the nature of the dispute, and any other important case characteristics.
• What are you trying to prove?—Outline your side of the argument and what you are trying to achieve in the case.
• Issues—Enter a list of issues and their descriptions. Issues should be in the following format: Issue name: Description.
• People and Aliases—Enter the names and aliases of key individuals in the case. Include their role and any other affiliations.
• Noteworthy Organizations—Enter the name of organizations and other relevant entities involved in the case. Highlight any key relationships between the entities or their notable characteristics.
• Additional Context—Enter any additional information that does not fit the other fields. This section is typically left blank.

Deposition outline and witness summary considerations

Deposition outline and witness summary creation uses Generative AI to read the facts and issues and then write a document of significant length from them. We strongly recommend that you provide an appropriate amount of oversight by having an attorney working on the case review the document, and confirm or correct their language. For further Generative AI considerations, see the Considerations regarding the use of Generative AI section below.

There is not currently a way to load prompt criteria from a prior run. We recommend that you use a text editor or word processor to preserve prompt criteria for future use.

Working with deposition outlines and witness summaries

Once a deposition outline or witness summary has been created, you can download copy to your workstation or create a saved search to share the documents that it contains with others. A deposition outline or witness summary can also be edited using the text editor in the Case Documents widget.

Downloading a deposition outline or witness summary

You can download a .docx version of a deposition outline or witness summary to your workstation from the text editor. The downloaded document is text-based and does not include any hyperlinks.

To download a deposition outline or witness summary:

1. Navigate to the Case Home tab and select the deposition outline or witness summary you wish to download in the Case Documents widget.
2. Click on the Download Word Document icon in the text editor toolbar.
   A copy of the deposition outline or witness summary is downloaded to your workstation.

Creating a saved search for a deposition outline or witness summary

The documents used in a deposition outline or witness summary can be shared via saved search to make it easy for the recipient to access the relevant documents. Once you've created a saved search, you can also export it using Import/Export. To learn more, see Exporting a saved search load file.

To create a saved search, add the Case Documents field as a condition and select the desired deposition outlines or witness summaries to include the relevant documents in the new saved search. To learn more about how to create a saved search, see Creating or editing a saved search.

Editing a deposition outline or witness summary

Once you have created a deposition outline or a witness summary, you can view its contents by expanding the Case Documents widget and then selecting the deposition outline or witness summary in the widget list. You can then edit the document with the text editor.

After making an edit, a message denoting the unsaved changes displays in the editor and the Save button is available. Once you are finished editing, click Save. If you have unsaved changes and try to navigate away from the case document, a pop up displays. Click OK to discard any unsaved edits and navigate away from the current document or click Cancel to stay on the current case document so that you can save your edits.

Alternatively, you can enable autosave to save any edits you make every 15 seconds. After each autosave, the timestamp next to the autosave toggle updates to help you track when the last autosave occurred. With this option enabled, your changes are automatically saved when you navigate away from the current case document.

Document Preview panel

The Document Preview panel is available from the Case Home tab. You can use the Document Preview panel to view documents without having to open them in the Viewer. To learn more, see Document and item list navigation.

General writing guidelines

For the prompt criteria in all of the setup tabs, we recommend:

• Write as if less is more. Instead of pasting in a long review protocol as-is, summarize where possible and include only key passages.
• Phrase things in a positive way when possible. Avoid negatives such as not statements and double negatives.
• Do not include explanations of the law.
• Do not give the LLM commands, such as, “you will review XX." Instead, simply describe the case.
• Use whatever writing format makes the most sense to a human reader. For example, bullet points might be useful for the People and Aliases field, but paragraphs might make sense in another field.
• The LLM understands widely used slang and abbreviations, but it does not necessarily know jargon or phrases that are specific to an organization.

Size limitations

Based on the limits of the underlying LLM, aiR for Case Strategy has size limits for the documents and prompts you submit.

The documents and prompt criteria have the following size limits:

• Each Generate Facts field has an individual limit of 4,096 characters except for the Prompt Criteria Name field which has a limit of 256 characters.
• Each document's extracted text must be under 600 KB. Documents larger than 600 KB will not have facts generated because they exceed the limit of the LLM. Similarly, facts are not generated for documents with little or no extracted text.
• Each document's extracted text, when combined with the prompt criteria, must be less than 120,000 tokens which roughly corresponds to word count but this can be inflated for documents that contain a lot of numbers or punctuation.
• To ensure high performance, there is a limit of 5,000 documents at a time when generating facts.

Considerations regarding the use of Generative AI in aiR for Case Strategy

Generative AI technology, specifically large language models (LLMs) are central to the capabilities of aiR for Case Strategy. While powerful, these technologies do carry risk, and it’s important that you understand the extent of those limitations as you use the product.

Broadly speaking, we expect the users of aiR for Case Strategy to always have a human overseeing the work produced by the AI. In particular, the possibility exists that the AI can do the following:

• Omit something important, such as failing to extract an important fact from a key document.
• Write something incorrect about something important, such as giving a low fact score to a key fact, or describing some part of the case in a deposition outline inaccurately.
• Make something important up, such as creating a fact that is not present in the document.

While our testing has produced very little evidence of these behaviors, we cannot be certain that they will never occur, and we want you to act with appropriate care.

Some further system limitations are explained below in more detail.

Non-determinism

Generative AI varies somewhat in its response even when asked the same question. This will manifest in aiR for Case Strategy in a few ways:

• The number of facts generated from a document can differ from run to run.
• The wording of facts can differ from run to run.
• The score, rationale, and other descriptions related to the facts can differ from run to run.
• Witness summaries and deposition outlines will have differences if run multiple times, even if the input facts list and the prompt criteria are the same.

Possibility of hallucinations

aiR for Case Strategy uses techniques to try to reduce the probability of what are known as hallucinations. Hallucinations, for our purposes, are when the LLM outputs something which is not grounded in the document from which the output is generated. Although we have not witnessed this behavior in rigorous testing, these events are still possible. Here are some issues that could potentially occur:

• Facts that do not actually reflect what is in the document can potentially be created.
• Details of facts, such as the fact date, might differ from what is actually in the document.
• Witness summaries or deposition outlines could reference things which are not in the input documents or facts.

We strongly recommend that an attorney or administrator carefully review the outputs to verify their accuracy as part of your overall workflow.

Extracted text only

Facts are generated based on the extracted text of documents. Any information that is stored in metadata, images, or is otherwise not available to someone reading the extracted text of the document cannot be used to inform generated facts. We recommend that you avoid including documents with misleading or empty extracted text while generating facts.

Fact fields used for generating case documents

To generate witness summaries and deposition outlines, aiR for Case Strategy uses only the main fields for the facts submitted including description, issues, date, etc. and the document summaries for the documents linked to those facts. In particular, metadata and extracted text from documents, apart from the document summary, are not currently used to construct case documents.

No memory from document to document

A single document is sent to the AI for fact generation at a time. It is then processed using the instructions that Relativity provides and the information you specified in the prompt criteria. For this reason, information that is present in another document will not be used in the formation of a fact. For example, a document might only reference an early code name of a project which is relevant to the case. If there’s nothing in the prompt criteria defining that early code name as being equivalent to a later project name, then aiR for Case Strategy will not be able to make the connection. This could potentially cause aiR for Case Strategy look past this document even if the project itself is relevant to the case. Of course, if the description were similar enough, aiR for Case Strategy could catch it, but overall, it’s good to establish aliases like this in the prompt criteria that you provide.

Another effect of this behavior is that similar or near-duplicate input documents can yield duplicate facts because the LLM does not know that the same fact has already been found on another document.