Build Knowledge Graph Intents
FAQs
Swift Responses
to common user queries
Provide question-answer sets to relevant nodes in the hierarchy to deliver an intelligent FAQ experience to the users. Add synonyms and channel-specific responses.
Task
Instant Resolution
to complex user requests
Link dialog task to Knowledge Graph intent and provide an aha experience to your users by handling their FAQs involving complex conversations.
ONTOLOGY
Drive Involved Conversations
by generating knowledge graph
Construct a Knowledge Graph (KG) – a hierarchy with key domain terms, make it more natural by adding context-specific questions and their derivatives, synonyms, and ML models.
While working on large data sets, leverage the Kore.ai Knowledge Graph Generator from Github to generate an efficient knowledge graph.
TAXONOMY
Efficient Disambiguation
for accurate responses
In the real world scenario, users may provide incomplete or ambiguous inputs, which virtual assistants may fail to understand and provide incorrect responses.
Kore.ai’s advanced NLP capability in the form of a Taxonomy based approach and built-in flows help engage the user in a multi-turn conversation to clarify user inputs and identify the right intent.
ANNOTATION
Optimize Responses
by marking the critical parts of the content
Leverage the powerful Annotation tool to annotate documents identifying the key sections of the content. Mark them as Header, Heading, Footer, Exclude or even Ignore page. It will help the Knowledge Graph engine to extract content efficiently and process it to deliver optimized results.
Analysis
Deliver Best Experience
by reviewing the question-answer set
A careful analysis of the Knowledge Graph helps detect errors in your questions and the path associated with the same that hamper the user experience.
Knowledge Graph Diagnosis tool allows you to identify any inefficiencies in your knowledge graphs and suggest possible corrective actions.
Kore.ai Knowledge Extraction
Extract, review, train and publish FAQs from structured and unstructured data in minutes, with ZERO coding
Use Disparate Data Source
Extract information from unstructured content, including web pages, PDFs, images, documents as well as structured content in .csv or JSON formats.
Intelligent Extraction
The KG engine automatically detects tables of contents and bookmarks as well as PDFs with multi-column layouts and stores the information within as assistant responses.
Review Extracted Content
Easily verify extracted questions. Modify responses to ensure your virtual assistant responds correctly with corresponding dialogs.
Add to Knowledge Collection
Selectively add the extracted questions to the appropriate taxonomy nodes in the knowledge collection.
Train and Publish
Add synonyms and alternative questions to enrich the virtual assistant’s understanding of user intent.
Manage the FAQs
Disable/re-enable selected nodes or FAQs invalid for a certain period. Also, use the new API to synchronize the FAQ status from your CMS.
