InsightAI: Frequently Asked Questions

This article serves as a comprehensive resource to address common queries and provide clarity on the platform's features, functionality, and benefits. It assists users in navigating InsightAI effectively, resolving technical concerns, and maximizing its potential for productivity. This page combines concise explanations with actionable insights.

What is InsightAI?

InsightAI is a GenAI chatbot experience available on Arctera™ Insight eDiscovery website that can be used across eDiscovery search results for fast and efficient summaries. It is a conversational AI assistant to efficiently review and analyze your relevant data. It uses Azure OpenAI service behind the scenes, which provides REST API access to OpenAI's powerful language models including o1, o1-mini, GPT-4o, GPT-4o mini, GPT-4 Turbo with Vision, GPT-4, GPT-3.5-Turbo, and Embeddings model series.

What is Azure OpenAI?

InsightAI, powered by Azure OpenAI Service, ensures the highest standards of data privacy and security. Your interactions are fully protected, and your data remains exclusively yours. Here's how:

• Your inputs (prompts) and outputs (responses) are not shared with other customers.
• Your data is not accessible to OpenAI.
• Your data is not used to train, retrain, or improve OpenAI models.
• Your data is not used to enhance Microsoft or third-party products or services without your explicit consent.
• Fine-tuned Azure OpenAI models are available exclusively for your use.

    For more details, refer to Data, privacy, and security for Azure OpenAI Service - Azure AI services | Microsoft Learn.

How does InsightAI work?

InsightAI offers seamless interaction within Arctera Insight eDiscovery to enhance your search experience. Each prompt sent to the Large Language Model (LLM) includes:

• System instructions
• Document context
• Your question

InsightAI operates in two distinct ways:

• Individual Item Level:
  • After performing a search, preview an email or file and click the InsightAI icon to ask questions about the specific item, including its attachments or contents.
  • At the individual item level, context is built from the email's contents and attachments, or the entire content of the file.
• Search Level:
  • After performing a search, click the InsightAI icon to ask questions related to all items in the search results, spanning emails, files, and chats.
  • At the search level, context is created by analyzing and summarizing as many items as possible within a 128,000-token limit.

Token Usage Insights: (Note: Tokens are explained in the answer to the next question.)

• InsightAI predicts the number of tokens your prompt will consume in real-time if you click the Send button.
• Once the response is received, InsightAI displays the actual number of tokens used during the transaction for full transparency.

Important Disclaimer

• InsightAI provides AI-generated content designed to assist with general queries.
• While efforts are made to ensure relevancy and accuracy, the content may not align with the latest legal standards or offer personalized guidance.
• Always consult a qualified legal professional for legal matters.
• Do not use InsightAI to seek legal advice.

What Are Tokens and How Are They Counted?

Tokens are pieces of words used by OpenAI models to process and generate text. In this case, when you send a request to InsightAI, your input is divided into tokens, which may include sub-words, spaces, or even partial words.

A thumb rule for tokens:

• ~1 token ≈ 4 characters in English
• ~1 token ≈ ¾ of a word
• ~100 tokens ≈ 75 words

Examples:

• 1-2 sentences ≈ 30 tokens
• 1 paragraph ≈ 100 tokens
• ~1,500 words ≈ 2,048 tokens

This framework helps users estimate token usage for effective interaction with language models.

How to enable InsightAI? 

To enable InsightAI at the customer level, log in to the Arctera Insight Management Console as Administrator and follow these steps: 

Enabling InsightAI at the customer level:

1. Log in to the Arctera Insight Management Console.
2. Navigate to Policy Management > Archive Options.
3. Find the InsightAI section at the bottom and toggle the switch to enable it.

Enabling InsightAI at the user level:

1. Go to Role Management > Assign Accounts.
2. Search for the desired reviewer and select their account.
3. In the Role Change section, enable the InsightAI privilege.
4. Click the Save button.
5. Repeat these steps for all reviewers you want to grant access to InsightAI.

What is the limit on free token usage?

When you enable InsightAI, a new token wallet is automatically created, and 1 million (1,000,000) free tokens are deposited for you to try the service. Once all tokens are exhausted, InsightAI will be disabled for all reviewers until additional tokens are added.

How to monitor token usage in InsightAI?

In the Arctera Insight Management Console, do the following:

• To view recent withdrawals, deposits, and top withdrawers, select Dashboard > Token Wallet.
• To view a trend report of token consumption, select Reports and Notifications > Usage > Token Trend.
• To view a detailed statement of all transactions, select Reports and Notifications > Usage > Token Wallet Statement.

What to do if all the tokens are exhausted?

To increase the limit, you need to contact Arctera Support. Refer to this article for detailed instructions on how to reach us.

Which OpenAI model InsightAI is using?

InsightAI currently uses the GPT-4o mini model (version 2024-07-18), which is the latest. It is a small, fast, cost-effective, and highly capable model. For more details, refer to Azure OpenAI Service models - Azure OpenAI | Microsoft Learn.

Where does InsightAI send my data?

InsightAI currently uses the following Azure OpenAI deployment types to send your data:

• Global Standard: Prompts and responses can be processed in any geography where the relevant Azure OpenAI model is deployed.
• Data Zone Standard: Prompts and responses are restricted to processing within a specified data zone, as defined by Microsoft.

The table below provides an overview of the deployment types used in various regions by the Arctera Insight Platform:

For more details, refer to:

• Understanding Azure OpenAI Service deployment types - Azure AI services | Microsoft Learn
• Data Residency in Azure | Microsoft Azure

Is My Data Being Logged?

InsightAI and the Arctera Insight Platform do not log any data used by it. However, Microsoft logs content sent to the Completions API and the Image Generations API, along with the responses to detect and mitigate harmful use of the Azure OpenAI Service. If flagged by the service's filters, the content may be reviewed by a full-time Microsoft employee.

For more details, refer to:

• Azure OpenAI Service abuse monitoring - Azure OpenAI | Microsoft Learn
• Azure OpenAI Service content filtering - Azure OpenAI | Microsoft Learn

Where to Learn More About Azure OpenAI

For additional information about Azure OpenAI, refer to:

• Azure OpenAI Service documentation | Microsoft Learn
• Code of Conduct for Microsoft Generative AI Services | Microsoft Learn
• Microsoft Responsible AI | Microsoft AI
• Data, privacy, and security for Azure OpenAI Service - Azure AI services | Microsoft Learn
• Data Residency in Azure | Microsoft Azure