Transforming Customer Insights Into Organizational Learning Via...

Transforming Customer Insights Into Organizational Learning Via Artificial Intelligence

Carina Ahlberg, Head of Customer Experience, Volkswagen Group Sverige

Carina Ahlberg, Head of Customer Experience, Volkswagen Group Sverige

I have been working with customer satisfaction and customer experience at Volkswagen Group Sverige AB for 15 years. We are the national sales company for Volkswagen, Volkswagen Commercial Vehicles, Audi, SKODA, SEAT, Porsche, and now also CUPRA in Sweden. I work closely together with all brands, our headquarters, and dealers. Our journey has gone from struggling with low interest in the surveys and customer feedback, via reward systems for customer satisfaction, to developing intelligent systems for the voice of the customer.

I would like to highlight some challenges that we are facing within the area of customer experience:

1. As part of a global organization, we need a KPI that can be followed over continents and time, e.g., the 5-star rating.

2. We are traditionally organized in silos.

3. The CX-tools are mainly management tools.

The 5-star rating gives us a way to compare with ourselves and within the organization. My experience is that an organization that strives for 5-star ratings tend to forget the person behind the stars, the customer. It gives no information about what the customer really thinks and feels in the different interactions with us.

Being organized in silos might be a good solution when it comes to internal processes, it gives us a way to handle everyday work, and we can keep knowledge updated within our area of competence. However, it is less efficient in terms of taking care of customers, since they have their own journey regardless of how we are organized. Can we continue to make customers adjust to our organization, or would it be more profitable and smart for us to adjust to the customer journey?

My third point refers to the tools we use to present findings regarding customers. In a world that needs quick feedback and action, these tools need to be easily accessible and structured in a way that helps the reader to learn from it. As an example, if I am developing a tool for booking workshop visits online, I need customer feedback for this specific touchpoint. In our current platform, we have plenty of feedback, but it is displayed per feedback, per brand, per organizational unit. In the end, we don’t have time to find or use it to develop the workshop booking online tool even though the customers give us very useful feedback.

While doing our job with customer journeys, it was clear that we had retrieved new, valuable insights. Now we could understand the behavior and underlying emotions that drove the customers in their decision-making. The reason we could trust the findings is that experts in human behavior and linguistics had analyzed the customer feedback. Now we are getting closer to why I am writing this article. We got the idea that the same people that analyzed the open interviews done for the customer journeys could train a machine to code all available customer feedback in written form and present it in an attractive way.

We have had a version of text mining in our current CX-platform. In my opinion, we have had some problems with both the coding and the presentation of it. In the worst case scenario, our Swedish customer comments have been translated into English, coded, and presented within the Swedish framework. At times, it has been really confusing since the automated translation hasn’t been able to understand more than words, and a language always words in a context. To make the coding trustworthy, we need to work with Swedish experts in the area. Since 2018 we have also had NLP, Natural Language Processing, in Swedish, which opens up a new arena for machine learning.

I also mentioned the presentation of the coded customer feedback. Our internal processes decided how to define the code-plan and how to display the coded comments. Now the code-plan will be decided by the customers, from the learning of what is important to them, which we can see when reading the comments and interviews. What are the customers actually talking about? What do they want to tell us? Moreover, with what emotions do they express their opinion?

With new transfer learning technique as BERT, Bidirectional Encoder Representation for Transformers, that is based on NLP we will now be able to automate the coding of open comments from all written sources, such as surveys, Google rating, e-mails, and social media.

We will present customer insights, including emotions, online and in line with the customer journey. All developers in our organization will have access to the platform in order to learn from it and come to quicker and more precise decisions. The new technique will be an important step to overcome our challenges. It will help us to speed up the transformation into becoming the truly customer-centric organization that we need to be. That is why I am writing this article.

Weekly Brief

Top 10 Brand Management Agency - 2021
<

Read Also

The New Reality of Loyalty

The New Reality of Loyalty

Claudia Infante, Vice President of Revenue Strategy at Margaritaville
The Most Known 5 Secrets to Generate Value with CRM

The Most Known 5 Secrets to Generate Value with CRM

Marco Ottaviani, Head of CRM and Customer Activation at MSC Cruises
Modern Tech Leaders Are Building Apps, Not Buying Them

Modern Tech Leaders Are Building Apps, Not Buying Them

Mike Pehl, Managing Partner, Guidepost Growth Equity
How Clubhouse is creating a better healthcare

How Clubhouse is creating a better healthcare

Florent Edouard, SVP, Global Head of Commercial Excellence at Grunenthal Group
How to Measure the Success of Third-Party Solutions on Your Website?

How to Measure the Success of Third-Party Solutions on Your Website?

Michael Goldrich, Global Head of Digital Marketing, Club Quarters Hotels
The Boom in Unicorn Startups isn't Going Anywhere

The Boom in Unicorn Startups isn't Going Anywhere

Audrius Milukas, Partner, Open Circle Capital