how DNX helped IndeCX using genAI

Ensuring a positive customer experience ranks high among business priorities, directly influencing profitability and growth. IndeCX, a platform specialising in customer experience monitoring and NPS (net promoter score), highlights that dissatisfaction can prompt as much as 78% of customers to switch to competitors, marking it as the leading cause for brand defection. 

Notably, IndeCX insights also reveal that 42% of customers are willing to pay a premium for a warmer, more welcoming experience, while about 72% share positive brand experiences with six or more people. Enhancing customer experience can yield significant financial rewards, with potential revenue increases ranging from 10 to 15%.

In this case study, DNX introduces the Generative Artificial Intelligence (GenAI) project for IndeCX, aimed at furthering their commitment to understanding and enhancing customer journeys.  Through this initiative, the company seeks to solidify its position as a leader in providing strategic solutions tailored to help businesses monitor and analyse customer experiences effectively. 

ABOUT INDECX

IndeCX helps businesses, mainly in the aviation and retail sectors, improve their customer relationships. With a 14-year track record in satisfaction and loyalty research, it operates across 30 countries and is recognised as a premier NPS® tool.

Using data-driven methods, the company extracts valuable insights through artificial intelligence to analyse large volumes of comments and feedback. IndeCX’s focus on customer experience (CX) alongside text data summarisation and SWOT analysis, provides a distinctive advantage in aiding product, service, and business strategy improvements.

THE BUSINESS CHALLENGE: A Scalable Summarisation and SWOT Analysis Solution

IndeCX’s core operations revolve around data analysis and deriving valuable insights for its clients. However, the company faced a significant challenge in strategically and efficiently analysing tens of thousands of comments.

Manual operations at IndeCX proved unfeasible: the insights derived could be imprecise, potentially harming product development, consumer satisfaction, and companies’ formulation of effective strategies. This risked customer attrition and impeded the company’s growth trajectory. Given the substantial volume of textual comment data, a scalable solution for summarisation and SWOT analysis became imperative.

In response, IndeCX turned to artificial intelligence and big data solutions to automate its processes. The objective was to enhance accuracy and value in analyses, thereby providing high-level support for clients’ decision-making.

The Solution: Text Summarisation and SWOT Analysis with Amazon Bedrock

To solve IndeCX’s challenge in extracting, processing, and analysing comments with AI and Big Data, DNX proposed a comprehensive solution with the most modern GenAI technologies and AWS cloud computing services.

The project involved creating an application using the LangChain framework to generate automated responses to user comments, which were then stored in Amazon S3 for access.

To summarise the large volume of data obtained, techniques such as MapReduce (a concept based on Big Data) and Refine were applied so that the information could then be processed by Generative AI. This way, it was possible to scale the system for this high volume of comments without losing the quality of the solution’s response.

Amazon Bedrock and Foundation Model (FM) Claude 2 from Anthropic

 

Highlighted in this project, DNX used Amazon Bedrock, a highly advanced and comprehensive GenAI service from AWS, along with the Foundation Model (FM) Claude 2 from Anthropic, a leader in artificial intelligence quality. Prompt Engineering techniques were also implemented to enhance the quality of responses generated by the language model in both the summarization and SWOT analysis stages. Additionally, the RAGAS evaluation framework was added to generate performance indicators at the response quality level through the Answer Relevance metric, ensuring a process of continuous improvement.

The integration with the IndeCX system was done through an API with complete infrastructure in a development and production environment created by DNX. AWS services such as Amazon SQS and DynamoDB were also integrated to ensure scalability and persistent data storage.

The developed solution allowed for the automation of comment processing at scale, using the most modern Generative AI technology from AWS, as well as robust Big Data strategies and APIs. The result was an exponential increase in response speed and quality, surpassing client expectations.

THE OUTCOME FOR INDECX

The GenAI solution devised by DNX for IndeCX yielded a notable increase in capacity and quality for extracting strategic insights from customer comments.

Automation and processing for tens of thousands of comments

 

DNX’s solution enabled the new IndeCX system to summarise approximately 20,000 comments within just 5 minutes and conduct SWOT analysis on the same volume of data in 4 minutes. This represented a significant enhancement in agility and efficiency compared to manual processes. The solution outperformed workflows employing competing language models (LLMs) from Amazon Bedrock.

 

High relevance of generated responses

 

Utilising the Claude 2 model, FM from Amazon Bedrock, the relevance (Answer Relevance) of both summarisation and SWOT analysis reached 85% and 52%, respectively, based on metrics provided by the RAGAS evaluation framework.

 

Successful integration and overcoming challenges

 

The solution’s integration was seamlessly achieved through the robust API developed by DNX. Moreover, precise cost control per API call was successfully implemented.

 

Productivity gain and scalability

 

The productivity gain and scalability achieved were crucial outcomes for IndeCX, particularly amidst the continuous growth in client data volumes. With DNX’s support, the swift and cost-effective implementation of this advanced GenAI solution bolstered the company’s value proposition for its clients, positioning it for increased competitiveness and leadership within its sector.

CONCLUSION: Successfully Implemented GenAI Elevates IndeCX’s Value Proposition

Generative AI, particularly leveraging Amazon Bedrock from AWS, facilitated a transformative partnership between DNX and IndeCX, enhancing the value of its tools.

The establishment of dependable and efficient workflows for data summarisation and SWOT analysis had a considerable impact, both economically and in terms of relevance, on IndeCX operations.

The seamless integration of the solution with the existing application surpassed project expectations, demonstrating DNX’s commitment to delivering exceptional outcomes.