Success Story

AI-powered feedback analysis tool

  • AdTech
  • AI transformation
  • eCommerce
  • Healthcare

We crafted TreviseAI, a customer sentiment analysis tool, which helps to transform reviews into actions, get customized responses, monitor critical sentiment shifts, and discover actionable insights. It offers a universal solution tailored to specific needs in various industries, whether you’re: 

  • a retailer trying to boost buyer experience on your e-platform, 
  • a hospitality chain striving to enhance guest engagement, or 
  • a healthcare provider seeking to improve patient satisfaction.
Link to project

Project numbers

2
months

took us to release the MVP version

3,000
characters

        may be processed           at a time

5
teammates

were involved in the dev process

Project details

Expertise used:

NLP (natural language processing)

Team composition:

  • Backend Engineer
  • Frontend Engineer
  • Manual QA Engineer
  • DevOps Engineer
  • Project Manager/Biz Analyst

Service provided:

  • Generative AI
  • Software and data architecture
  • UI/UX design
  • Custom software development
  • MVP building
  • Manual QA

Project background

We, at GreenM, have noticed lately an abundance of reviews, comments, and social media posts, which impact purchasing decisions and attitudes towards products and services.

 

However, manual methods of analyzing customer sentiment proved time-consuming and lacked scalability.

 

This inspired us to create a sophisticated sentiment analysis tool aimed at addressing the growing need for effortless and effective customer engagement strategies.

Technology challenges

  • Interface adaptation: transforming the tool into a chat-like interface without significant code modifications required prompt engineering;
  • Output consistency: providing the user with consistent and human-readable output, despite variations in the language model’s responses;
  • Emoji encoding: Ensuring accurate text encoding and proper display of emojis in the output to enhance user understanding and experience;
  • Prompt protection: Preventing the language model from deviating from the instructions and ensuring compliance with user requirements. 

Tech stack

Python

Business challenges

Helping businesses deal with endless customer reviews and extensive online presence

Analyzing customer feedback manually often falls short in providing timely and insights

Processing and replying to feedback written in a language other than service provider’s one

Solution delivered

By leveraging expertise in data engineering, advanced Generative AI, and NLP, the team behind Trevise has developed an AI-driven sentiment analysis tool intended to revolutionize how businesses interpret and respond to customer sentiment and which:

  • analyzes text data accurately, even dealing with sarcasm, irony, and other creative expressions without getting confused; 
  • makes sense of unstructured feedback data by categorizing it as positive, negative, or neutral and by respective service in one click; 
  • identifies critical sentiment shifts, enabling proactive interventions to address emerging issues and open new opportunities; 
  • suggests possible ways to improve the service or product based on the complaints received to enhance customer satisfaction.

Key features & capabilities

  • Quick reply suggestions: you can respond to your customers in a friendly manner that aligns with their feedback, providing an exact reaction to their words; 
  • Category-based results: you will see compliments, criticism, or neutral comments highlighted with a respective color and service category in outcome; 
  • Service alerts: you can stay ahead with announcements of critical sentiment shifts in your customer feedback, even if it appears to be soft in tone;
  • Actionable intelligence: your customer feedback will be analyzed effortlessly and transformed into actionable strategies through service improvement suggestions; 
  • Multilingual support: you can deliver top-notch service to all your customers by analyzing feedback in multiple languages, ensuring you are on the same page; 
  • Ease-of-use and accessibility: The tool was designed to be user-friendly, and it is available as the website platform and ChatGPT interface.

Results were reached through

  • Prompt engineering approach: An iterative prompt engineering approach was employed to align the sentiment analysis tool with chat interfaces seamlessly. This approach involved refining instructions and providing examples to guide the language model in generating accurate and human-readable responses. 
  • Few-shot learning approach: Text encoding issues and emoji display inconsistencies were addressed through the use of few-shot learning approaches. By providing examples within the prompt, the model was trained to accurately interpret text encoding and display emojis correctly in the output.
  • Prompt protection mechanisms: Additional penalties were introduced to the model to enforce prompt protection, ensuring that the language model adhered to the specified task and user instructions without deviation.

Business values

Efficient data processing

Our sentiment analysis tool automates the process of analyzing large volumes of textual data, enabling organizations to process information quickly and efficiently in one click.

Strategic decision making

Tool provides valuable data-driven insights into customer preferences, pain points, and satisfaction levels, which can inform product development, marketing strategies, and customer service improvements.

Competitive advantage

By staying informed about customer sentiment and market trends, you can gain a competitive edge by identifying opportunities, mitigating risks, and differentiating yourself from competitors

Enhanced customer experience

Sentiment analysis helps you identify areas for improvement in products or services, enabling them to address customer issues proactively and deliver a better overall customer experience.

Scalability and cost-effectiveness

Feedback analysis tool is scalable and can handle increasing volumes of data efficiently, making them cost-effective solutions for businesses of all sizes. 

Business risk mitigation

Sentiment analysis can help identify potential risks and threats to a business, such as negative public perception, customer dissatisfaction, or emerging market challenges.

For whom?

Our platform is ideal for any businesses, from startups to multinational corporations, seeking to gain actionable insights from client feedback and drive meaningful improvements in their operations. 

Learn our clients’ experience

More case studies

Online property bidding platform 

View more →
We created a property auction product from scratch with only client's idea. Its goal was to streamline the property buying and selling process, make it intuitive, scalable, and transparent to all users.
  • MVP development
  • Real estate
  • 200 req/sec can be handled by new bidding website
  • 4 mos took to get MVP out + first agents enrolled
  • up to 1,000 users can use the site concurrently

Copyright © 2024 GreenM, Inc. All rights reserved.