How Sentimate’s AI technology
turns data into insights.

How does Sentimate turn raw data into insights?
Using proprietary AI algorithms, of course.

From data…

Data Collection
Collect unstructured opinions from consumers, using publicly available sources, such as eCommerce websites

Data Cleansing
Remove fraudulent, duplicated, or promoted opinions and ensure the quality of the extracted data

Research any product

…to insights

Data Structuring

Organize and structure the data

Data Classification
Classify data by category, brand, and product variations, and identify topics of discussion and perform sentiment analysis

Quality Assurance
Ensure the insights are accurate and well organized

From data to insights

Workflow

The workflow behind Sentimate’s insight includes five main steps: data collection, cleansing, structuring, classification, and quality assurance. Each of these steps is performed using a variety of proprietary and non-proprietary algorithms meant to ensure that the insights we offer you are accurate and reliable.

workflow

Artificial Intelligence

Applying common AI algorithms to perform these tasks — from data collection to classification — would be virtually impossible: it would take so much time, so many people, and so much money to train and maintain the technology that it would not be worth it. Sentimate relies on proprietary and unique artificial intelligence (AI) that does require human intervention and training for the definition of topics and keywords, ensuring speed and accuracy.

Artificial intelligence

Topic Extraction

Consumer opinions are nuanced and often address multiple issues related to the product within the same sentence. For example, a consumer might say: “I love the design, but was disappointed with the price.” Our algorithms extract from every sentence the topics discussed in it without the need to predefine in advance which topics to extract. These include common topics such as size, design, price, and ease of use, as well as category-specific topics, such as noise cancellation for headphones, navigation for robot vacuums, and wetness indicator for diapers.

Topic extraction

Sentiment Analysis

Generic sentiment analysis engines such as BERT only help up to a certain degree of accuracy. Unlike previous tools, Sentimate’s engine is able to analyze opinions in relation to the product’s specific domain. For example, a sentence like “The battery lasts 1 hour” is positive for a drone but negative for a laptop; or the sentence “The toast came out brown on both sides” is positive in the specific context of bread toasters.

Topic extraction

Et voilà!

Sentimate users don’t need to know anything about algorithms and artificial intelligence. The solution we created is for everyone to use, from R&D to marketing and finance professionals. The Insights Dashboard gives instant access to well-structured and easy-to-read product insights, such as SWOT analysis and competitive landscape. Whether you need a quick snapshot or an ultra-granular analysis, Sentimate has got you covered.

Et voilà!

Ready to start using Sentimate?

Create your own user feedback survey