The startup brings sentiment analysis to non-technical users

by Max Smolaks 7 August 2019

American startup Luminoso has made a number of major tweaks to Daylight, its natural language processing application.

Daylight’s primary purpose is to analyse customer feedback – hidden in reams of unstructured text across online artefacts like product reviews, open-ended survey responses, and support tickets.

The latest version introduces advanced sentiment analysis, helping users understand the feelings associated with brands, products and services, across more than a dozen languages – and the system takes emojis into account too.

There’s also a new insights dashboard called Highlights, that enables employees in marketing, sales or product development to upload feedback datasets related to a specific product or service, and immediately find out which issues affect customer rating the most, or what the major themes of customer conversations are.

Luminoso has been working on machine learning tools for sentiment analysis since 2010. The company is headquartered in Cambridge, MA – the home of the Massachusetts Institute of Technology (MIT), where its founders originally developed their algorithms.

Luminoso’s largest customers include Sony, Autodesk, Scotts Miracle-Gro, and GlaxoSmithKline.

The aim of Daylight is to enable less technical users to automate the feedback review process, without having to write any code. Like humans, the software learns words from the context in which they are written, rather than dry dictionary entries. This means Daylight can understand words and phrases it has never seen before.

“We built Daylight to offer all users — including those without data science backgrounds — the fastest, most intuitive way to automate the understanding of customer feedback datasets without barriers, such as the need for training machine learning models and creating ontologies,” said Ying Chen, chief product officer at Luminoso.