Kyndi’s Reading Automation Engines
The Kyndi Reading Automation platform automates a range of reading tasks that people usually perform manually – including scanning, skimming, intensive reading, and prioritizing what they read. This dramatically accelerates the reading process, while retaining human-like levels of interpretation.
The platform is comprised of the following AI engines and tools:
Kyndi’s Discovery Engine
Kyndi’s Relevance Engine
Kyndi’s Explanation Engine
Kyndi’s Lexicon Engines
Scan or skim a series of documents with the Discovery Engine to identify frequently mentioned keywords and phrases to understand what’s important and detect trends. Kyndi’s powerful data cleansing methods detect and remove noise and non-natural language from the text in a fully automated, unsupervised manner—including unnecessary formatting, punctuation, and elements such as headers, footers, and confidentiality statements.
Then Kyndi transforms large volumes of unstructured text into a powerful AI knowledge base that contains extracted and inferred information about concepts, topics, entities, and relationships within the content. Kyndi’s AI-powered Discovery Engine is highly accurate and efficient, providing users with the ability to do things like:
- Identify what keywords (e.g. product names, nouns, locations, people, email addresses etc.), custom entities or phrases are most prominent.
- Quantify how often these topics are mentioned.
- Create a dashboard to detect trends in topics and quantities in documents analyzed over time.
When users need to find hidden intelligence in large volumes of text, it’s essential to use natural language processing to query. Kyndi’s unique ability to analyze and detect not just keywords – but also a host of concepts related to the search term – surfaces all related ideas in the long-form text, allowing users to find greater quantities of more relevant data. That’s how the Kyndi Relevance Engine enables exceptional querying accuracy and speed to insight.
The Relevance Engine helps users to:
- Search for a phrase or specific term and identify documents that include that content to find answers quickly (e.g. querying wiki files for the answer to a support question.)
- Drill down to identify the most relevant information. Users can perform an initial search, then ask additional questions about the returned data to identify the resources that hold the most relevant intelligence, saving valuable time.
When users find new intelligence, they may need to understand the context in which it was used and its specific location. Kyndi Explanation Engine makes it fast and easy for users to dig in to see surrounding text around result terms. Users can also identify the content’s specific location and filename so they can quickly access the exact document they need or keep a record of the source.
Reading automation platforms perform better when they understand relevant terms. Every industry uses names and vocabularies that have specific meanings. Kyndi Lexicon Engines let users import lexicons so the Kyndi platform can instantly recognize terminology unique to their business or industry, speeding results and improving output quality.
Important documents can be found in a wide range of locations. Kyndi provides users with a range of connectors and APIs to make it easy to read and write to filestores, databases, and content management systems. Kyndi allows content loading to common knowledge bases such as Amazon S3 and can work with customers to set up new connectors. And Kyndi reads un-labelled, unstructured text in formats such as PDF, DOCX, HTML, TXT, etc.
Virtual Private Cloud (Cloud Agnostic) or On-Premise using enabling technologies like Elasticsearch/Kubernetes/Docker for rapid and flexible deployment