Surety is performing system maintenance this weekend. Electronic date stamps on new Prior Art Database disclosures may be delayed.
Browse Prior Art Database

Context based question auto-completion and suggested answers for messaging

IP.com Disclosure Number: IPCOM000249464D
Publication Date: 2017-Feb-28
Document File: 6 page(s) / 321K

Publishing Venue

The IP.com Prior Art Database


This publication details a system for automated retrieval of answers to questions in a group chat thread in the case that the information already exists. This would mean users would not have to search for the information in a message thread as it is readily presented, and in the case that the user is unaware of the existence of the already existing information, alert them to the fact it exists.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 26% of the total text.

Context based question auto-completion and suggested answers for messaging

Most people use some form of instant messaging including texting and various social media messaging services. Conversations through these platforms range from private 1-1 message threads to large group chats between friends, colleagues and other social groupings.

In group chats, due to the number of people resulting in a high volume of messages which consist of spam, personal conversations, old topic continuations etc, it is easy for people to miss or overlook a specific piece of information (e.g. the time and place the group are to arrive at an event). This is made worse if an individual is not following the conversation at the time, and misses the information. This results in one or many of the group having to ask / re-ask for that piece of information, and can be frustrating for those providing details as well as for those who cannot find the details.

In 1-1 (and group) conversations, it is possible that the pair (group) have a long history and someone cannot remember a specific piece of information, or if they have even received that information in the past (e.g. address of house of longstanding friend provided a long time ago). This results in the pain of having to ask for that piece of information again, which requires the other person to see the message and respond, which if the need is urgent and the other person does not have their device with them, can be a major pain point.

Currently there exists no simple solution to provide a user access to such past information. By exposing already provided answers and details to a user posing a question on a messaging platform, the frustration of repeating information can be avoided, and productivity on the part of the asker can be increased. System Overview

The system is able to detect when the user is asking a question and suggests auto-complete endings to the question for which the system is aware that information already exists in the message thread.

Once the user has selected a question suggestion or completed their own question, the system uses natural language processing to analyze and evaluate the question. It then searches through the message thread to find relevant information that has already been provided. Once evidence has been collated, it is presented to the user using natural language as a suggested answer to their question.

This is advantageous over prior art as it streamlines the process of finding existing information (eg. rather than asking to be provided information, just to be told that it has already been said, so you then have to search and sift through options), and can also pick up answers to questions that the user did not realize already existed, thus saving the user the hassle of having to ask and then wait for a reply. System Details

The system consists of two main stages: the initial stage detects the user is asking a question and attempts to help formulate a question by finding previously ...