AIFields use the power of artificial intelligence to extract and organize information from documents. AIFields use a file as input and extract various types of information or generate summaries from the file, based on a list of analysis types.
The extracted information in AIFields can be pulled into reports or referenced in generated documents, just as other fields can be.
Note: This feature requires Docs+ and incurs a usage-based charge. Reach out to your account manager to learn more.
How AIFields Work
An AIField is an additional option for a multi-line text field, added in the Customs Editor. This text field is connected to a single file attachment field, and the desired analysis type is selected.
When a file is uploaded to the connected file attachment field, the AI analysis will begin. This process usually takes a few minutes. Once it is finished, the information, determined by the selected analysis type, appears in the text field.
Analysis Types
Below is the list of all available analysis types with an explanation of what each does.
| Document Type | Analysis Type | Description |
|---|---|---|
| Accident reconstruction report | Summarize the conclusion and key findings | Extract the main conclusion and key findings. |
| Answer to RFP | What is opposing counsel worried about? | Analyze requests for production and identify the main issues or themes that the opposing counsel seems most concerned about. |
| Clinical research | List negative findings about a product | Identify negative impacts associated with a product. |
| Contracts | List contract terms | Extract key terms, including dates, payments, and amounts due, in a list. |
| Contracts | Identify important terms and clauses of the contract | Extract key terms like start and end date, amount due, and jurisdiction. This type also extracts any important clauses, such as indemnity, non-compete, arbitration, etc. |
| Contracts | Identify potential dispute clauses | List out clauses that may become contentious in the future, including any terms that may be ambiguous, unfair, or potentially going against any existing regulations. |
| Deposition | Identify any inconsistencies in the deponents deposition testimony | Extract all logical inconsistencies and provide the page and line numbers associated with them in a bulleted list. |
| Deposition | Summarize the client depo and identify any red flags | Provide a summary and list any statements that could potentially be problematic or need further attention. |
| Deposition | Summarize the deposition | Provide a summary of the deposition along with key themes. |
| Draft Discovery Responses | Identify potential objections raised by the opposing party | Review and predict potential issues or objections that may be raised by the opposition. |
| Summarize the email chain | Analyze and convert a threaded email into natural, readable language. | |
| Identify client questions in need of a response | Analyze the text within an email chain or letter and look for inquiries or questions from the client that warrant a response. | |
| Identify action items | Analyze an email chain, identify any outstanding tasks or actions, and summarize the current state of the matter. | |
| Identify potential areas for negotiation | Review email exchanges between parties to indicate possible negotiation points or areas of compromise that could be beneficial during settlement discussions. | |
| Expert Report or CV | Summarize the expert report or CV with a scientific literature focus | Summarize an expert witness report or CV with a focus on extracting data that relies on published scientific literature. |
| Expert Report or CV | Provide a summary of the expert witness CV -with a professional history, academia and publications focus. | Summarize an expert witness report or CV with a focus on scientific publications, and their role in academia and in any regulated bodies. |
| Expert Report or CV | Provide a list of sample deposition questions based on an expert report | Provide a list of starter or sample depo questions to help draft an early deposition outline. |
| General Legal Response (Depo, Answer etc) | Create an objection log | Analyze and create a log of all objections contained. If there are page or line numbers, those will be included. |
| Independent Medical Examination (IME) | Summarize the IME/DME report | Create a summary of the attached report from an IME/DME and distill it into simple terms. |
| Insurance declarations page | Extract key policy information | Extract coverage limits, policy numbers, type of policy, and other key information. |
| Interrogatories | Summarize the plaintiff's employment history | Scan an interrogatory and provide a summary of the employment history of the plaintiff. |
| Interrogatories | Identify prior legal actions the defendant has been involved in | Identify any prior legal proceedings the subject has been involved in. |
| Lease agreement | Identify clauses related to eviction | Extract the specific clauses related to eviction in a lease agreement. |
| Lease agreement | List terms and conditions | Extract the key terms and conditions of a lease agreement. |
| Medical records | Medical chronology | Review a set of medical records and for each treatment date will return: Date of service, Provider, Chief Complaint, Pain Score, ICD-10-CM Codes, CPT Codes, Diagnosis, Treatment, Medications, Follow ups or referrals. |
| Medical records | Identify all injuries, treatment and surgeries described in the records | Extract and list all injuries, treatments, and surgeries mentioned in the medical records. |
| Medical records | Pain scores over time | Extract all pain scores mentioned in the medical records and present them in an ordered format, showing the progression of the patient's pain over time. |
| Medical records | Extract CPT codes | Identify and list all Current Procedural Terminology (CPT) codes found within the medical records, which describe medical, surgical, and diagnostic procedures. |
| Medical records | Detail by treatment date | Sort and present all treatments provided by date, creating a comprehensive timeline of the patient's medical journey. |
| Medical records | Sample deposition questions | Generate a set of sample deposition questions that could be asked during a deposition, based on the information within the medical records. |
| Medical records | Create Safety Rules | Analyze the medical records and generate a list of rules which could have prevented injury, if adhered to. |
| Medical records | Extract ICD-10-CM Codes | Identify and list all International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) codes found within the medical records, which represent medical diagnoses. |
| Medical records | List the medications prescribed to the client | Identify any medications prescribed to the client, as well as their dosage. |
| Medical records | Identify future treatment | Extract mentions of planned or suggested future treatments, including all treatments suggested by healthcare professionals, their frequency and duration, and any specific conditions or precautions associated with them. |
| Medical or scientific literature | Summarize and fact check the journal article | Provide a summary and confirm the facts and data mentioned within it. |
| Police report | Summarize what happened | Provide a concise summary, including parties involved, date, location, and events as described in the report. |
| Police report | Identify areas of ambiguity | Identify areas of ambiguity or that the opposing party might question. This may include unclear timelines, witness statements, or general descriptions. |
| Requests for Production (RFP) | Based on the requests in this document, what is the requestor worried about | Identify the main concerns or priorities of the requesting party. |
| Trust | Summarize the trustee's duties | Summarize the duties imposed on the trustee. |
| Will | Identify all beneficiaries | Review a will, identifying and listing all the named beneficiaries. |
Set Up an AIField
Setting up an AIField requires access to the Customs Editor Advanced tool, which can be granted in the Advanced Access tab.
Navigate to the Customs Editor by clicking the main menu and selecting Advanced from the dropdown. Then, select Customs Editor from the left side menu. If prompted, select your Org. Select the project template where you want to add the AIField.
File Attachment Field
Before you create the AIField, determine which file attachment field you will be connecting to the AIField. You can use an existing file attachment field, or create a new one. Both single file attachment fields and multiple file attachment fields can be used.
The file attachment field can be in any static section, and does not need to be in the same section as the AIField. However, if either the file attachment field or the AIField is in a collection section, then the two fields must be in the same section.
One file attachment field can be used for multiple AIFields.
Add the Text AIField
Once you’ve determined your desired file attachment field, you can add the AIField.
Click the section where you’d like to add the AIField. An AIField can be in any static section. If the AIField is in a collection section, then the file attachment field must be in the same section.
Click Add Field/Widget, and select either the Text - Multiple Lines or the Text Multiple Lines (Large) field type. (Learn more about text fields.) Fill out the field name and other properties as you would normally.
Toggle on AI-Analysis. Once this option is turned on, you will see fields for linked file attachment field and analysis type. Select the file attachment field that you’ve chosen, and then select desired analysis type. To see a more detailed list of what each analysis type does, refer to the table above.
Once you’ve selected the file attachment field and the analysis type, click Create. The AIField is now ready to use! When that section is open in the Customs Editor, the AIField will be noted with an AI Linked badge.
AIFields in Project
You can now see the AIField in your project. Both the link file attachment fields and the text output fields will have a sparkle icon indicator to show that they are AIFields or AI linked.
When a file is uploaded to the connected file attachment field and the section is saved, the AI analysis begins. While the analysis is being geneated, you will see a loading icon. Mouse over the icon to see that status of the analysis. Once the AI analysis is finished for the document, this icon will become a green checkmark to show that the AI analysis is successful.
- A status of OCR in progress means the document is being processed and its text is being extracted.
- A status of AI Analysis in progress means the document has completed OCR and is now being analyzed by the AI.
- Once the job is complete, you will see a success status message. At this time, you can refresh your page to see results.
The time it takes for AI analysis to run on a given document can vary. Most already-OCRed documents are analyzed in a matter of minutes. However, this can take longer depending on current usage and document length.
When the AIField text has been generated and the page has been refreshed, it will appear in the text field with the header text “AIField Analysis Start,” the attached file name, and the date and time the analysis was generated. The end of the AI analysis for that document is marked “AIField Analysis End.”
The text can be edited at any point. If you delete the attached file, and reupload a new file, the generated text will be appended to the text field, rather than overwriting the existing text.
Multiple Uploaded Files
AI analysis is performed on every file that is uploaded to the linked file attachment field. If more than one file is uploaded to the file attachment field, the documents are analyzed one by one and the AI analysis for the newly uploaded file is added to the bottom of the output field, separated by the header and details.
Reporting
Admins may want to see how many pages of files AIFields have reviewed, understand how many AIFields they have generated over a set period of time, or discern how many times a certain analysis type was used.
AIFields information can be reported on using the new “AI Field Job List” report. You can find this report in the Create New Report list in the report builder.
The AI Field Job List report includes the following columns:
- Org name
- project name
- analysis type: the selected analysis type for this field
- filename: the name of the file the AI field drew from
- page count: the page count of the file the AI field drew from
- job status: whether the job is in Analyzing, Completed, Created, Queued, or Failed status
- username: the username of the user who uploaded the file
- email: the email of the user who uploaded the file
- target field: the name of the AI field
- target section: the name of the custom section containing the AI field
- created date: the date and time that the file was uploaded, when the AI field was triggered
- completed date: the date and time that the AI field finished generating
The AI Field Job List report can be limited by time period for either the created or completed dates, by project or Org, or by the following criteria concerning the AIField Job itself:
- analysis type
- file pages, including range
- user information
- status
Comments
0 comments
Article is closed for comments.