---
title: "Batch Analysis"
canonical: "https://help.agiloft.com/space/HELP/43714360/Batch+Analysis"
format: markdown
---
With the batch analysis add-on, you can:

- Train your own labels with [AI Trainer](https://agiloft-public.atlassian.net/wiki/pages/createpage.action?spaceKey=help&title=Creating%20Custom%20Labels%20with%20AI%20Trainer%20HELP&linkCreation=true&fromPageId=43714360)
- [Import legacy contracts](https://agiloft-public.atlassian.net/wiki/spaces/HELP/pages/43719546) to Agiloft and convert them to records, all with the click of a button
- Extract and review large batches of[ third-party contract data](https://wiki.agiloft.com/display/HELP/Using+the+Agiloft+Contract+Assistant+for+Word#Working-In-The-App) for comparison and negotiation

> ℹ️ **Prerequisites**
> ℹ️ 
> ℹ️ Batch analysis is an add-on.

# Deployment  


Work with your Agiloft representative to enable, deploy, and configure batch analysis. This process involves deploying the platform, setting up OCR Cloud services, and updating the KB with additional tables that are used for platform features. Training and extraction is only available in English.

# Tables

Deploying batch analysis automatically creates the following tables in your KB:

- **AI Projects**: This table holds information for two kinds of projects, listed below. This is the process table for AI Platform, where most workflows occur.
  - **AI Label Training**: used to create new labels for AI to use in contract import- and review-related workflows.
  - **Legacy Import**: used to process multiple contracts into records at once, using an AI model to extract relevant data. AI Label Training is used to widen the amount of labels that a Legacy Import project can extract automatically.
- **AI Project Documents**: This table holds contract attachments that are used to either train new AI Label Models or create new contract records. <span style="color: #333333">Usually, these attachments are uploaded in a related table in an AI Project record, rather than being added directly to this table.</span>
- **Label Library**: This table holds records that reflect general categories an AI Label Model may fall under, like Termination clauses.
- **AI Label Models**: This table is a subtable of Label Library, and holds a record for each individual label. An example of an AI Label Model record would be a Termination clause used specifically for Employment Agreements.
- **AI Model Versions**: This table holds a record for each iteration of a custom AI Label Model. This is useful if you want to use a version of the model that was trained using an earlier data set.
- **AI Document Types**: This table holds records of different kinds of contract documents. These are used to match AI Label Models with the contracts they've been trained to work on.

# OCR Language Support

Agiloft offers two kinds of OCR engine, [AWS Textract](https://aws.amazon.com/textract/faqs/#topic-0) and [Tesseract](https://github.com/tesseract-ocr/tesseract). For information on Screens language support, visit [Screens](https://agiloft-public.atlassian.net/wiki/spaces/DT/pages/43719552).

Textract is an AWS product that is optimized for English, yet also works with languages that use a Latin-based alphabet. It does not officially support non-Latin alphabets such as Arabic, Chinese, Japanese, Korean, etc. Tesseract, however, can be used for languages that do not use Latin-based alphabet. Tesseract is an open-source software that was initially developed by Hewlett-Packard, but is now run by Google.

> ⚠️ In the case that a language is supported by both Textract and Tesseract, users should opt for Textract. Textract typically provides better accuracy, without the need for additional configuration or pre/post-processing. It also returns a structured JSON output that can help your system process data.

## Textract

The following languages are supported by AWS Textract:

- English
- Spanish
- German
- French
- Italian
- Portuguese

## Tesseract

For use with any languages that are not listed above, Tesseract is the recommended option, It supports over 100 languages, with a full list located [here](https://tesseract-ocr.github.io/tessdoc/Data-Files-in-different-versions.html). Tesseract includes support for the following alphabets:

- Latin
- Greek
- Cyrillic
- Arabic
- Hebrew
- Indic
- CJK (Chinese, Japanese, Korean)
- Southeast Asian
- Ethiopic

#### Related articles

> Macro (contentbylabel)