Do you want to streamline your organisation’s Learning & Development, saying goodbye to classrooms & webinars?

Ingests information from audio files and stores it so that it can interrogated by numerous Soffos APIs.

Given a body of text, a selection of emotions to detect (e.g., frustration, excitement, worry, neutral) and instructions on how to segment the text, this module matches each segment to one of the given emotions. This can be useful for tracking customer satisfaction and provides more insights than simple sentiment classification.

Given a body of text and a question, returns a succinct natural language answer to the question. It can answer complicated questions, identify ambiguity and offensiveness and respond appropriately to random inputs such as greetings.

Ingests information from Video files (audio + description of visuals) and stores it so that it can interrogated by numerous Soffos APIs.

Processes various types of files (PDF, DOCX, TXT) to extract categorized pieces of text. This aids in the normalization of documents in a unified format in preparation for document ingestion.

Sometimes Speech-to-Text (STT) systems make mistakes when converting voice to words, especially when the speaker is not speaking clearly, or has a heavy accent. Transcript correction takes the output of those systems, finds those errors, and corrects them to what was most likely uttered, or meant.

Given a truth string and one or more comparison strings, the String Similarity module measures the level of similarity between the truth and each one of the comparison strings. A similarity score (0-100%) is returned, as well as spans of text which are highly similar between the two strings. This is useful for scoring user's answers to questions.

The TestMe™  app is a quick, efficient and fun way for learners in the education and/or corporate fields to get up to speed with essential knowledge in no time - and it's free.

Typically used in conjunction with the File Converter, ingests a body of text by preparing it to be interrogated by the rest of our modules such as Question Answering, Natural Language Search, etc.

Suggests tags for a body of text.  Useful for determining keywords and subjects.

AMA allows any customer to ask questions about a product,
an advertisement, or a webpage. As a user you will then receive actionable business intelligence including insights into what customers really care about, which supports sales and lead generation.

‍

Given a body of text, it returns snippets of text that are offensive or contain profanity.

Detects the language of a body of text. Supports more than 100 languages.

A tool that allows patients to schedule and manage their medical appointments via WhatsApp. AI determines the probability of no-shows and recommends action to reduce and mitigate them

Given a body of text, returns text which contains ambiguities. Includes many types of ambiguities, including those arising from word sense and pronoun issues. It, optionally, also returns questions that should be asked of a user to resolve the ambiguities.

Once content is ingested, this module returns contradictions to the user, thus offering the opportunity to resolve them immediately.

Generates a list of questions and answers from a selected body of text. Declarative statements are extracted from the text and rephrased into a question, where the answer is the original statement. This is useful for testing users' recall of specific information.

The natural language search module allows you to search through thousands of ingested documents by just asking a question and returns relevant passages with answers to your question and their original documents in a matter of seconds.

‍

Given a body of text, identifies and labels entities, such as people, places, etc.

This module summarizes a body of text to the specified length. It is able to blend facts together an formulate a summary "in its own words".

Enables a conversation regarding a selection of text. Learns from interactions with users and remembers context. Can also return questions when context is not clear.

Given a selection of text, keywords, or subjects, returns a ‘microlesson’, which is approximately three minutes’ worth of speech.

Given a selection of text and a ‘how-to’ question, returns a body of text that contains instructions on how to perform a process, containing statements such as "if-then", "repeat step", "go to step", etc.

Paraphrases a body of text according to the level of complexity required. Simplifying the text alters the wording so that it is better understood by a user, without losing any information or changing the meaning.

Detects illogical statements in a body of text and provides explanations as to why they are illogical.

Given a sequence of tags for a specific user relating to a specific set of documents, compares it to the sequence of tags of other users for that same set of documents and suggest adjacent knowledge.