Browse Prior Art Database

Method for Gathering Anonymous Feedback by Audio Analysis of Conversations in Central/Key Locations

IP.com Disclosure Number: IPCOM000247902D
Publication Date: 2016-Oct-11
Document File: 3 page(s) / 68K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method for gathering anonymous feedback through audio analysis of conversations captured in central/key locations known by participants.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

Method for Gathering Anonymous Feedback by Audio Analysis of Conversations in Central/Key Locations

A method is needed that enables companies to harvest the thoughts, feedback, and creative ideas that come from employees during informal and spontaneous conversations.

The novel contribution is a method for gathering anonymous feedback through audio analysis of conversations captured in central/key locations known by participants. The method records conversations that take place in known recording areas throughout a workplace. The system uses voice distortion to maintain anonymity. Data mining of the voice recording provides the company with candid feedback and access to ideas. This method of authentic but anonymous feedback can potentially replace existing 'office employee feedback surveys' and other formal attempts to capture employee input.

The main steps of the method include:


1. Install microphones in known and well-marked public areas so; people who carry out meetings in those areas are

well aware that their talks might be recorded

2. Each participant can opt to distort his/her voice and ask mask identity to remain anonymous 3. Use cognitive and machine learning de-personification post process to remove names (of those who want it) and other clear markers that allow identification of the participants 4. Harvest recording data and send to post-processing where systems harvest ideas, opinions, and themes of informal conversations

Optionally, post-processed information can be tagged with a participant's name before applying voice distortion or removing identifying characteristics. Participants could opt to review asso...