Revisions - QnABot on AWS

Revisions

Date Change
September 2021 Release v5.0.0 – Initial AWS Solutions Implementation release. For more information, see the CHANGELOG.md file in the GitHub repository.
October 2021 Release v5.0.1 – Bug fix for redaction of PII in logs; documentation addition for deploying a web UI. For more information, see the CHANGELOG.md file in the GitHub repository.
December 2021 Release v5.1.0 – Integration with Genesys. Fixed integration with Slack and LexV2. Intelligent PII redaction with Amazon Comprehend. Bug fix for Amazon Kendra FAQ and metadata tags for questions. Added client filter support to allow same questions answered differently based on session attributes. For more information, see the CHANGELOG.md file in the GitHub repository.
February 2022 Release v5.1.1: Expanded language support for voice and text interactions, included support for Neural voices for Amazon Lex language locales, fixed Amazon Kendra WebCrawler data source sync issues. For more information, see the CHANGELOG.md file in the GitHub repository.
March 2022 Release v5.1.2: Added logic to support Amazon Connect Interactive Messages and a new set of example questions to be imported for Genesys Cloud CX. For more information, see the CHANGELOG.md file in the GitHub repository.
July 2022 Release v5.2.0: This release includes: early implementations of intent and slot matching and Canvas LMS integration; support for using custom domain names in QnABot on AWS content designer and client interfaces; Command Line Interface (CLI) for QnABot; Amazon Kendra Redirect capability; ability to import QnABot questions and answers from an Excel file uploaded to an S3 data folder; support for importing session attributes using Excel files; bugs fixes and updates. For more information, see the CHANGELOG.md file in the GitHub repository.
September 2022 Release v5.2.1: This release includes security patches, changes for the AWS Lambda release that supports the Node.js 16 runtime, and a bug fix for the error caused by not providing an image URL in the Bot’s response card. For more information, see the CHANGELOG.md file in the GitHub repository.
October 2022

Release v5.2.2: This release includes npm and pip security patches; improved deployment stability for OpenSearch and Amazon Lex resources creation; single character utterance bug fix; and ElicitResponse bug fix. For more information, see the CHANGELOG.md file in the GitHub repository.

November 2022 Release v5.2.3: This release includes npm and pip security patches. For more information, see the CHANGELOG.md file in the GitHub repository.
November 2022 Release v5.2.4: This release includes npm and pip security patches. For more information, see the CHANGELOG.md file in the GitHub repository.
December 2022

Release v5.2.5: This release includes npm and pip security patches; documentation improvements; new unit tests; clientfilter bug fix; Amazon Kendra FAQ bug fix; missing Fulfillment Lambda function widget fix; and support has been added for the latest LexV2 languages. For more information, see the CHANGELOG.md file in the GitHub repository.

January 2023 Release v5.2.6: This release includes npm and pip security patches. For more information, see the CHANGELOG.md file in the GitHub repository.
February 2023 Release v5.2.7: This release includes npm and pip security patches; and new unit tests. For more information, see the CHANGELOG.md file in the GitHub repository.
February 2023 Release v5.3.0: This release moves the solution onto OpenSearch v1.3 and introduces new QnA search capabilities using text embeddings. By enabling text embeddings, users can leverage LLMs to obtain semantic-based query matching. For more information, see the CHANGELOG.md file in the GitHub repository.
March 2023

Release v5.3.1: This release includes npm and pip security patches; and a bug fix for theFulfillment Lambda function not correctly publishing a new version. For more information, see the CHANGELOG.md file in the GitHub repository.

April 2023

Release v5.3.2: This release includes npm and pip security patches; a bug fix for Alexa skill reprompts; new CloudFormation parameter to configure EBS volume size for OpenSearch; MetricsBucket Amazon S3 bucket added to CF output; updates to Amazon Lex and Amazon Connect response limits; and miscellaneous documentation updates. For more information, see the CHANGELOG.md file in the GitHub repository.

April 2023 Release v5.3.3: This release includes npm security patches. For more information, see the CHANGELOG.md file in the GitHub repository.
May 2023 Release v5.3.4: This release includes npm and pip security patches and a bug fix for Amazon Connect voice responses. For more information, see the CHANGELOG.md file in the GitHub repository.
July 2023

Release v5.3.5: This release includes pip security patches and removal of the ElasticSearchUpdate custom resource to prevent CFNLambda recursion alert. For more information, see the CHANGELOG.md file in the GitHub repository.

July 2023 Release v5.4.0: This release introduces additional QnA search capabilities using LLMs. By enabling LLMs end users can leverage additional features such as query disambiguation, text generation to answer questions from an Amazon Kendra index or from the new text passage item type. This release also updates the Lambda Runtimes to Nodejs18 and Python 3.10 and adds initial support for AppRegistry integration. For more information, see the CHANGELOG.md file in the GitHub repository.
August 2023 Release v5.4.1: This release includes minor documentation updates to the LLM README and additional LLM guidance in the Implementation Guide. For more information, see the CHANGELOG.md file in the GitHub repository.
September 2023 Release v5.4.2: This release includes minor updates and bug fixes. For more information, see the CHANGELOG.md file in the GitHub repository.
October 2023

Release v5.4.3: This release includes fix for an issue where Alexa schema was not exporting the utterances list. and additional documentation on bot routing configuration.

Documentation additions include data storage and protection, guidance section for implementing quizzes, PII Redactions, Multi-language support & Bot Routing. For more information, see the CHANGELOG.md file in the GitHub repository.

October 2023 Release v5.4.4: Updated package versions to resolve security vulnerabilities. For more information, see the CHANGELOG.md file in the GitHub repository.
November 2023 Release v5.4.5: Updated package versions to resolve security vulnerabilities. For more information, see the CHANGELOG.md file in the GitHub repository.
January 2024 Release v5.5.0: This release introduces core-language parameter to the QnABot deployment that supports 33 Languages and allows the user to select a core language in which the OpenSearch language analyzers will be used. This provides a more syntactical accuracy for matching questions and answers without resorting to translation. This release also has additional enhancements like Bot routing, protected utterances settings, functional test collection and improvements in error handling. This release also has updates like Bluebird migration to native promises, upgrade to AWS SDK for JavaScript v3, Wepack 5, Vue3 and Vuetify3. It also includes documentation updates, code quality improvements, security patches and bug fixes. For more information, see the CHANGELOG.md file in the GitHub repository.
April 2024

Release v5.5.1: This release fixes Document Chaining issues, updates the QnABot Client from using Cognito Auth Code instead of Implicit Grant, and includes patched vulnerabilities as a part of these changes. For more information, see the CHANGELOG.md file in the GitHub repository.

April 2024

Release v5.5.1: This release fixes document chaining issues, updates the QnABot client from using Cognito Auth Code instead of Implicit Grant, and includes patched vulnerabilities as a part of these changes. For more information, refer to the CHANGELOG.md file in the GitHub repository.

June 2024 Release v6.0.0: This release introduces native integration with Amazon Bedrock, which provides access to the latest LLMs from leading AI enterprises (Amazon’s Titan, Anthropic’s Claude 3, Cohere’s Command, Meta’s Llama 3, Mistal AI’s Large Model) to find a model best suited for your use case. Additionally, this release also integrates with Amazon Bedrock knowledge base so you can retrieve specific, relevant data from your data sources, stored in Amazon S3 and automatically converted to text embeddings stored in a vector database of your choice (end-to-end managed RAG). You can then retrieve your company specific information with source attribution (for example, citations) to improve transparency and minimize hallucinations. This release also includes seamless switching between specialty bots, UI improvements, and other fixes. For more information, see the CHANGELOG.md file in the GitHub repository.