Skip to main content

Jane - Data Agent for business teams

Jane is our maiden data agent built exclusively for business teams to converse with enterprise data using natural language (without having to know SQL or Semantic language).

Jane leverages all the data assets and applications within Xceed Analytics and Generative AI to get the accurate answers to the user. Business teams can now truly self-serve, get answers to their business queries and/or create dashboards on their own instantly without much of deep data expertise or technical knowhow.

Data agent applications follow the same stringent governance and controls as rest of Xceed Analytics subsystem including table and row level access controls. Data Organisations can now truly enable business users to get answers to most common queries and also monitor and track the consumption and usage.

High Level Block Diagram

A high level architecture for Xceed Data Agent Application is as below

Jane - High Level Architecture

It consists of following components:

  1. Xceed Agent User Interface: This component enables business users to ask questions of data in a natural language and get insights from the enterprise data.

  2. Xceed Data Agent Application: This component provides the service interface to the actual agent functionality that brings Jane to life.

Xceed Agent User Interface

Xceed Agent has the familiar conversational experience for carrying out end to end tasks. The conversational view has two sub-panes:

Xceed Data Agent User Interface

  1. On the left is conversation or chat view, which further includes the conversation view on the top and the user input at the bottom
  2. On the right is the view of various user conversation sessions.
  3. The conversation view is further enhance to detect visualizations/reports which can be directly pinned to Xceed Workspace/Dashboard, right from the conversational interface

Jane from inside

Data teams can deploy Jane similar to any other workloads via Compute Environments in Xceed Analytics. They can instantiate Jane by department/function or as may be the case enabling distribution of queries and workload.

Jane like other agents in Xceed leverage Xceed Intelligence Framework which provides common building blocks for RAG and Agent application building. It further adds the following components to put the data agent together.

  1. Planner task: This task understands the user query and reasons the same and returns A DAG with the set of available Tasks for appropriately responding to the user query.

  2. Data Tasks: These are granular tasks which shoulder various responsibilities: These tasks can perform one or more of the following:

    • Identifying right set of datasets for a given task
    • Exploring metadata and operational data for the task,
    • Running statistical profiling tools to understand column level details
    • Constructing and running a semantic query and returning the insights as result
    • Selecting and plotting appropriate visualization for the insight
    • Finally, interpreting the insight and creating a user summary for the insight.
  3. Tools Integration:

    • Integrates with all other Xceed Analytics tools
    • Xceed Vector store for identifying right dataset
    • Xceed Data catalog for trusted examples
    • Xceed SQL querier and Semantic Layer for executing the query
    • Xceed Pipeline builder for data transformation and cleanup

Get the power of futuristic Data & AI Platform for your enterprise.