Working with Context Grounding

This section includes information on how to use Context Grounding effectively.

Getting started

To use Context Grounding with agents or Autopilot for Everyone, create an index following the steps described in Creating indexes.

To use Context Grounding with activities, create a connection to the UiPath GenAI activities connector and make sure you use Studio Web or Studio Desktop version 2024.4 or newer.

Managing the ingestion pipeline

You can manage the ingestion pipelines through:

• Orchestrator, from the Indexes page. See Managing indexes.
• The Update Context Grounding Index activity, part of the UiPath GenAI activities package.

Querying data with Context Grounding

After creating an index in Orchestrator, indexes are accessible throughout the UiPath platform. These indexes serve as persistent storage for documents ingested from your data sources, offering a reusable resource for various UiPath products:

• In Autopilot for Everyone, Context Grounding enhances user interactions by enabling searches across existing indexes to provide accurate answers to queries. For details, refer to Context Grounding in Autopilot for Everyone.
• GenAI Activities benefit from Context Grounding by allowing content generation based on information stored in permissioned knowledge bases. For details, refer to GenAI Activities.
• For agents, indexes play a crucial role in providing context during runs. For details, refer to Contexts.

Monitoring Context Grounding

Understanding how Context Grounding influences your workflows is crucial for optimizing performance and troubleshooting. Here's how you can trace and view Context Grounding outputs across different UiPath products.

In Agents, access the Trace view of the agent run to see comprehensive details. This view provides all search results and citations from the Context Grounding query, offering insights into the agent's decision-making process.

To gather detailed information about Context Grounding in GenAI activities:

1. Place a Log Message activity immediately after the Content Generation activity in your workflow sequence.
2. Use the following output variables to capture specific information:
   • Top Generated Text: View the LLM generation response after workflow execution.
   • Citations: Examine the semantic search results that influenced the generation output. This works only for PDF and JSON data types.

Context Grounding in GenAI Activities

Context Grounding interacts with your data in three phases:

1. Establish your data source for Context Grounding.
   • Context Grounding follows shared folder permissions. Use a folder with appropriate access to manage and query from data.
   • Create a connection to the supported Integration Service data sources or add data to a shared Orchestrator bucket location.
2. Ingest data from your data source into Context Grounding.
   • Create and sync indexes in Orchestrator.
   • Use the Update Context Grounding Index activity to sync an index you have created.
3. Query and ground prompts with your data.
   • Use the Content Generation activity, agents, or Autopilot for Everyone to query over documents and use information to augment or ground prompts.

Common Context Grounding patterns

The core components of Context Grounding are designed to provide a mechanism that supports finding pertinent information within and across documents, and surfacing only the most relevant pieces needed for a high-quality, low-latency generation from an LLM.

Searching within documents

The Context Grounding service helps you find specific information within a single document more effectively. Instead of just matching keywords, it understands the meaning and context of your search query. For example, if you're looking for information about "apple pie recipes" in a cookbook, it would understand that you're interested in desserts and baking, not technology or fruit farming.

Searching across documents

Context Grounding helps you find information spread across multiple documents. It can understand the relationships between different pieces of information and provide more relevant results. For example, if you're researching "climate change effects on agriculture" across various scientific papers, it pulls together relevant information from multiple sources, understanding that topics like rainfall patterns, crop yields, and temperature changes are all related to your query.

This means you can use Context Grounding for:

• Data extraction and comparison: Context Grounding can automatically identify and extract specific types of information from documents, then compare them in meaningful ways. Imagine you have a stack of résumés and want to compare candidates' work experiences. The service could extract job titles, durations, and responsibilities, then present them in a way that makes comparison easy, even if the information is formatted differently in each résumé.
• Summarization: Context Grounding can create summaries of long documents or multiple related documents. It doesn't pick out random sentences, but understands the key points and overall message. For example, if you have a long report on market trends, the service can provide a summary highlighting the main findings, key statistics, and overall conclusions.

Notifications

You can subscribe to receive notifications from Context Grounding. Visit Notifications panel to learn more.

Events serve as triggers for notifications. The Context Grounding events that generate notifications are:

• Ingestion Job Completed
• Ingestion Job Failed
• Ingestion Job Started

You can also subscribe to events based on their severity, such as Success or Error.