Skip to main content
Skip table of contents

Release Notes - 4.0

Intended audience: ALL USERS

AO Platform: 4.0

Release Date

  • March 2024

Overview

  • Release 4.0 builds on previous 3.x releases of the AO Platform. We continue to introduce many new features that will make the development and management of solutions much more effortless and be done mainly from the browser as a low-code environment.

  • The 4.0 release continues to have a strong focus on UI and functional enhancements to the AO Easy Answers solution as well as the composers supporting Easy Answers directly, including Ontology/MSO Composer, Solution Composer, and Model Composer.

  • This topic includes a high-level introduction to the enhancements made during this release.

New in Release

AO Easy Answers

  • New Insights Feed page where all Insights will be shown in chronological order. See Easy Answers - Insights Page.

  • New and updated Quick Insights. See Quick Insights Based on Machine Learning Models.

    • Updated Time-series Insights - single time-series insights (Outlier, Change Point, and Trend/Seasonal Trend Detection), - are now all based on Bayesian ML Models.

    • New Forecasting Insight for single time-series data.

    • New Comparison Insights - for categorical data, the following new ML Models have been introduced:

      • Comparison of Two Categories (Binomial Analysis) - The binomial analysis compares the observed counts of two samples of data and determines the proportion that they are different.

      • Means Comparison (T-Test Analysis) - The t-test analysis is used to determine if there is a significant difference between the means of two groups. Two options are available for One-Sample and Two-Sample T-Tests.

      • Comparison of Distributions (Chi-Squared Analysis) - The chi-squared analysis examines a distribution of proportions and discovers the high and low points in the distribution.

      • Correlation Insight.

    • New Multivariate Time Series Insights, including…

      • Global Multivariate Trend Detection, Seasonal Trend, and Correlation Insights.

      • Multivariate Change Point Detection.

  • New “Did you mean…?” suggestions if a Natural Language Question is not fully or partially understood using new Vector Store database integration. See Easy Answers Error Messages.

    • Includes option to “Fix Unknown Words” and much enhanced Synonym creation with full Annotate workflow - replacing the simpler Annotate option in Edit Query dialog.

  • New Customize App option in Results page of Easy Answers - allowing users to quickly change an App visualization to an alternative (compatible) visualization, including Preview, eg. change a Donut Chart to a Column Chart, or a Column Chart to a Waffle Chart. See Customize Apps. New Charts introduced in this release include:

    • Funnel Chart

    • Half Pie / Half Donut Charts

    • Polar Chart

    • Pyramid Chart

    • Packed Bubble Chart

    • Radial Chart

    • Waffle Chart

  • Enhancements to Downloading of Data for Apps and Insights. See Downloading Data.

    • Downloading data as a background task for large data volumes.

    • Option to download multiple data tables as either zipped CSV files or as sheets in an Excel file.

    • Option to select/deselect which columns to be included in the download.

    • Option to download from History of downloads.

  • New Enable Advanced Filter option when Adding an App to a Dashboard. This allows the user to add additional filters using the Query Builder to an App for easier quick filtering of the data represented in the App.

Chatbot

  • The Chatbot integrated with Easy Answers has seen some significant improvements.

  • Search in Content Repositories. See Search in Content Repositories

    • Ability to upload documents (Word, PowerPoint, and PDF) to a content repository.

    • Ability to ask Natural Language Questions (NLQs) to retrieve text from a document repository, eg. service manuals, contracts/agreements, HR policies, or any other repository of unstructured data documents.

    • This functionality uses integration with a Large Language Model (LLM), such as OpenAI.

  • Conversation threads. See https://apporchid.atlassian.net/wiki/spaces/AOPHC/pages/2701623297/Easy+Answers+ChatBot+Conversation.

    • Ability to ask NLQs as a threaded conversation with the Chatbot. In other words, multiple questions can be asked in sequence within which the context is preserved allowing questions to be additive until a new conversation is started.

    • Open a popup with Dashboard for any question in the conversation.

    • Open questions from the History of conversation threads.

    • Get question suggestions from Chatbot if the user question isn’t fully or partially understood.

    • This functionality uses integration with a Large Language Model (LLM), such as OpenAI.

  • General UI and functional enhancements, including a new Chatbot icon, the addition of a microphone icon in the question field allowing users to speak a question, and improved Chatbot prompting with suggestions in case of unknown words in questions.

Ontology Composer

  • Support for Apache Druid, Trino, Microsoft SQL Server, Microsoft SQL Server Analysis Services, Azure Synapse Analytics, and Elasticsearch as JDBC database data sources in Ontology Discovery. See Discovering an Ontology.

  • New Ontology Composer (similar to MSO Composer) provides new configuration options at Ontology level. See Editing an Ontology.

    • Content Repositories page - allowing users to create one or more Content Repositories that can be used as a resource for questions in the Easy Answers Chatbot. Also see Chatbot section above.

    • Settings page - includes many new configuration options for enabling the use of Generative AI LLMs for both Summarizations of Data and Quick Insights as well as code generation (Text2SQL) during Natural Language Processing. Also, new configuration options for enabling the use of a vector database, including Weaviate or Pinecone, for generating alternative question suggestions in Easy Answers.

  • More MSO Expressions - used to create new Derived MSO Properties and for more advanced MSO Curations. See Using Expressions, MSO - Properties, and MSO - Curation.

  • MSOs can now be saved as Versions and restored from Version History if required. See MSO - Version History

  • The configuration of the Key field for an MSO has been moved from the Properties page (checkbox) to the General page (repeater section of field properties). If more than one field property is configured, a composite Key is created from the field properties specified.

  • Numerous updates to MSO Settings in MSO Composer, including…

    • Easy Answers page > Search Store options for Token Matching.

    • Insights page > options for configuration of Textual Insight Templates.

Insight Composer

  • New functionality allowing users to create Custom Models (using Pipeline or Java Class for input). See Insight Composer Page.

Pipeline Composer

  • Support for two new vector databases, Weaviate and Pinecone as both data source and data sink. See Editing a Pipeline.

  • Support for Apache Druid, Trino, Microsoft SQL Server, Microsoft SQL Server Analysis Services, and Elasticsearch as JDBC database data sources.

Admin Solution

  • A new page for Easy Answers Questions Audit Analysis was added to the Reports section, allowing the administrator to get statistics on question counts per user per solution. See Admin > Reports.

  • A new section of pages was added for Generative AI configurations, including Prompt Templates, LLM Configurations, Embedding Models, and LLM Tasks. See Admin > Generative AI.

  • A new section of pages for Ontology configurations relative to LLM Prompts, including MSO Word Prompts, MSO Prompts, and MSO Property Prompts.

Other UI Improvements

  • The option to use LLM in Easy Answers has been removed from the User Settings dialog. This option as well as many other new LLM-related configurations are now available in the Ontology Composer > Settings pages. See Editing an Ontology.

API

  • New API endpoints for working with integrated vector databases.

Infrastructure

  • Availability of Helm charts for the deployment of AO Platform in an auto-scaling Kubernetes cluster.

  • Support for MicroK8s for deployment of AO Platform in single-instance Kubernetes cluster. See Deployment of AO Platform on MicroK8s.

Libraries

  • Upgraded Highcharts from 7.2.2 to 11.1.0

  • Upgraded Python from 3.6 to 3.9

  • Upgraded Webix from 7.03 to 10.20

Documentation

  • Updated the Open Source Library list included with AO Platform and Easy Answers.

  • The AO Platform documentation can be accessed from https://docs.apporchid.com - or from options in the User menu in the AO Platform UI. This main user documentation is not publicly available, a token is required to access it unless accessed from within the AO Platform UI in the User menu.

Deprecations/Removals

  • None

System Requirements

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.

Contact Support Close