We are excited to announce the release of the Kore.ai Experience Platform Version 9.0. This release includes the following key features and enhancements:
Conversation Dialog Builder
The new Dialog Builder offers an innovative approach to build your Virtual Assistant’s skills. Design your use cases as conversational flows and the platform will intelligently transform them into Dialog Tasks. The conversation flow and the Dialog Task definition are always kept in sync so that the business teams can continue to refine the flow and the engineering teams can build the integrations. Know more
Automate your everyday business process like leave application, expense reimbursement, employee onboarding Building, etc. using the Process Apps. Design multi-step, multi-user workflows using a rich set of building blocks like integrations, business logic evaluation, conditional flows and approval flows. Integrate with Virtual Assistants for offering conversational and omnichannel experiences. Users can also use the self-service portal to create new requests, provide approvals and track the requests. Know more
The all New Interface
The Virtual Assistant designing features are now classified into Build, Deploy, Analyze and Manage sections, aligning them closely with the development lifecycle. All the features required to build tasks, train, and test are included under the Build menu. Channel enablement, integrations, publish flow, and bot definition management are under the Deploy menu. Know more
The store offers a wide range of templates to accelerate your development process. Explore the Store to find a template that meets your needs, install and start using it. Know more
This release includes support for 5 new languages - Catalan, Tamil, Hindi, Marathi, and Telugu
The Versions feature now allows you to visually compare the differences between two versions of a bot across various sections of the definition. Know more
NLU & NLP Features
- KAEN ML Model - The Kore Advanced Embedding Network (KAEN) ML Model allows your virtual assistants to precisely identify the user intents by understanding the meaning of a user sentence and at the same time give importance to the domain-specific terminology. Know more
- Unified NLP Training – A simplified experience for training the Machine Learning and Fundamental Meaning Engines in one place. Intents can be trained using a unified interface, making it easier to switch between utterances, patterns, and rules.
- Taxonomy-based Knowledge Graph – Explore the new approach to build a Knowledge Graph with built-in flows to engage the user in a multi-turn conversation to disambiguate the path and find the right question. Proceed to the Advanced NLP Configurations section to enable this feature. This is a beta release for this feature and a few more updates are expected in the upcoming releases.
- Bot training can be further refined by configuring the following additional settings for the NL engines:
- Custom configuration for auto spell correction support for ML predictions.
- Additional configurations for the evaluation of the Fundamental Machine Engine Patterns.
- Configure Softmax Temperature to define how confidently the ML Engine should identify the winning intent from the ML Model.
- Configure Neurons in Hidden Layer to determine the intensity/rigor to be adopted while performing intent identification by the ML Model.
- Entity Behavior
- Ability to accept only ‘month and year’ as input for Date entity (returnOnlyMonthYear)
- Ability to set preferred format for Date entity to avoid ambiguity scenarios (preferredDateFormat)
Administration and Security Enhancements
- The new pay-as-you-go plans provide flexible options to deploy virtual assistants. The first five virtual assistants of an account can be deployed with 5,000 free billing sessions.
- SAML enhancements:
- Option to map SAML attributes with various Roles and Groups in the Kore.ai account for automating the user authentication.
- Ability to add multiple SAML certificates to support fallback
- JWE support has been extended to the Webhook channel.
- Node Module has been upgraded to handle Security Vulnerability.
You can find the complete release notes on our Documentation page. See Release Notes