Conversing with Your Data

Intended audience: ALL USERS

AO Easy Answers for Mobile: 2.0

Overview

Easy Answers for Mobile solution provides a conversation-based approach to asking questions and getting answers from the solutions ontology. The Agent is based on an Agentic AI framework. The Agent combines the power of the AO Platform Ontology (providing the data and visualizations) with the advances in using Generative AI to both interpret the user questions as well as provide articulated Natural Language Responses and a visual representation of data for a much-improved user experience.

Video

Asking Questions

The following are examples of where questions can be asked in the Easy Answers solution. Questions are asked using a Natural Language approach and must generally relate to the business ontology used by the solution. It’s recommended to only ask questions that are naturally building on previous questions in the conversation.

As a question is typed into the NLQ field, the Agent automatically spell-checks each word and corrects misspellings in real time, showing corrected words underlined.

What Are Verified Questions?

Some questions in the Agent are marked with a blue "Verified" badge . This means an authorized user has reviewed the question and confirmed its results are accurate, so you can rely on them as-is. To keep that trust intact, Retry and LLM Feedback aren't needed for verified questions since the response has already been checked and signed off on. See details in Using Question Verification - Mobile.

On The Welcome Screen

The Welcome screen is shown with some question suggestions. For example, “show all meters“.
Tap a suggested question, or type other questions in the input field.

• Instead of typing the question, the user can tap the microphone icon (when the keyboard is visible) and use voice typing.

In a Conversation

On the conversation page of the Agent, additional questions can be entered in the Ask a follow-up question… box at the bottom of the page.

Follow-up Questions

Once a question has been answered, one or more suggested follow-up questions will be added to the bottom of the conversation page for easy selection and execution. Simply tap on one of these questions to continue the conversation.

From the History Panel

Questions are organized into conversations, so it’s recommended to only ask questions that are naturally building on previous questions in the conversation. If the user wants to ask questions that are not directly related to previous questions, tap the “+ New Chat” button in the Conversation History screen. The Conversation History screen accumulates all conversations in a History list with LLM-generated Titles and Descriptions.

The user can tap on any of these entries to resume a previous conversation, or swipe left on a conversation to delete it. See Loading a Conversation from History.

Processing a Question

Whenever a question is asked, the Agent will start the process of interpreting the question and generating a query that will be sent to the underlying database(s) in order to retrieve the data to answer the question. A process box appears on the screen showing the steps along the way, including the AI Reasoning when the process uses the LLM to interpret the question.

Answering a Question

Once ready, the Agent will provide an answer to the user’s question. An answer typically includes one or more of the following components:

• A short, concise textual response.

• A summary of some additional insights about the data.

• A visualization.

Textual Insights

Actions on a Textual Result

• Verified - icon - tap to show a tooltip with details about the verified question status. See Using Question Verification - Mobile.

• About - icon - see Questions Requiring Clarification(s).

• Thumbs Up - icon - tap to validate the results shown; a simple acknowledgement will be shown on the screen. This will signal to the Agent that the user is satisfied with the response, and the Agent will, over time, learn and optimize future responses.

• Thumbs Down - icon - tap to open a dialog where the user can provide some input as to why the thumbs-down icon was tapped. Provide as much information as possible. When tapping the Submit button, the Feedback will be sent for approval (or rejection) by a “supervisor” user with additional permissions to update the Ontology to benefit future users asking the same or similar questions.

• Retry - icon - tap to rerun the question to have the data in the response updated.

• Options Menu

  • AI Reasoning - tap to open a new screen with the “chain-of-thought” (or Reasoning) output from the LLM, allowing the user to understand how the LLM interpreted the question and the steps taken to respond.

Visualizations

When a question involves data that can be represented visually, the Agent automatically selects the most appropriate chart or table format and displays it alongside the textual response. Each visualization card includes a title, a brief description of what the chart represents, and an expand icon (↗) in the top-right corner to view it in full screen.

Actions on Visualization Results

Visualizations may have different actions. Here are the most common actions:

• Show As Table - - tap this option to turn data in a Chart into a table with the raw data.

  

  
  

• Change Chart Type - - tap to show a pop-up of alternative, compatible chart types. Select the chart type to use instead of the current chart.

  

  
  

• Pages - if a Table is shown as the visualization, a page selector will be shown at the bottom of the screen, allowing the user to view large volumes of data when the table content cannot be shown on a single screen.

  

See full list of Visualizations.

Contact App Orchid | Disclaimer