Improving Customer Experience with Artificial Intelligence, Data Analysis, and Automation in Azets
Tirkkonen, Terho (2021)
2021
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2021082517139
https://urn.fi/URN:NBN:fi:amk-2021082517139
Tiivistelmä
Net Promoter Score or System (NPS) is globally a leading management system for measuring and improving customer experience (CX). It is a straightforward metric and a powerful tool but using NPS for producing a competitive advantage requires data aggregation and analysis and a deep understanding of the nature of CX.
This development project is created for Azets to understand the predictive role of NPS in customer buying behavior. Azets has used the NPS methodology since 2013; as a result, the company has collected a vast amount of customer feedback data.
The primary purpose is to provide insights and proposals for improving customer experience and retention at Azets Finland and Azets Group. The objective is answered by analyzing NPS data with artificial intelligence (AI) and combining NPS data with invoicing and churn data, in addition to utilizing a comprehensive literature review when proposing development projects.
This thesis follows the single-case study research method. The research contains two parts and uses secondary data collected by Azets from small and medium-sized (SME) customers initially for other purposes than for the thesis.
The first part of the research combines quantitative NPS data, invoicing data, and churn data to seek connections between customer feedback and buying behavior. The second part of the research uses AI and natural language processing (NLP) to interpret 32,000 customer comments – collected in NPS surveys between 2013 and 2021 – to find key topics, sentiments, and feelings that affect customer experience at Azets.
Findings reveal the most valuable customers are not always Promoters with an NPS rating of 9 or 10 on a scale of 0-10 and that the NPS does not predict churn at Azets in the same way as described in theory. Qualitative data analysis with AI shows that the critical CX factors do not match the categories predefined in Azets’ NPS process.
The research findings and suggested automation projects can enable Azets to increase revenues by millions of euros with moderately small additional investments. Most of the appendixes and found insights contained confidential information and are not included in the published version. Such data removed from the public version include, for example, the average billing in each NPS category and the sentiments identified by artificial intelligence related to critical CX factors. The thesis was evaluated based on the non-public version.
This development project is created for Azets to understand the predictive role of NPS in customer buying behavior. Azets has used the NPS methodology since 2013; as a result, the company has collected a vast amount of customer feedback data.
The primary purpose is to provide insights and proposals for improving customer experience and retention at Azets Finland and Azets Group. The objective is answered by analyzing NPS data with artificial intelligence (AI) and combining NPS data with invoicing and churn data, in addition to utilizing a comprehensive literature review when proposing development projects.
This thesis follows the single-case study research method. The research contains two parts and uses secondary data collected by Azets from small and medium-sized (SME) customers initially for other purposes than for the thesis.
The first part of the research combines quantitative NPS data, invoicing data, and churn data to seek connections between customer feedback and buying behavior. The second part of the research uses AI and natural language processing (NLP) to interpret 32,000 customer comments – collected in NPS surveys between 2013 and 2021 – to find key topics, sentiments, and feelings that affect customer experience at Azets.
Findings reveal the most valuable customers are not always Promoters with an NPS rating of 9 or 10 on a scale of 0-10 and that the NPS does not predict churn at Azets in the same way as described in theory. Qualitative data analysis with AI shows that the critical CX factors do not match the categories predefined in Azets’ NPS process.
The research findings and suggested automation projects can enable Azets to increase revenues by millions of euros with moderately small additional investments. Most of the appendixes and found insights contained confidential information and are not included in the published version. Such data removed from the public version include, for example, the average billing in each NPS category and the sentiments identified by artificial intelligence related to critical CX factors. The thesis was evaluated based on the non-public version.