Customer Feedback Analytics‍_

• 40,000 of open-text feedback pieces / quarter

• 120 high-skill man-days spent on tagging the feedback

• 3 weeks needed to compile results after receiving a feedback batch

Albert’s Challenge

Each time after receiving a batch of NPS (net promoter score) feedback, Albert’s analysts sorted a fraction of it – around 30 % – into categories. The painful process was taking 120 man-days per year while not producing consistent results. Albert needed to speed it up and relieve its employees.

SentiSquare’s Solution

SentiSquare AI created a semantic model by learning to understand the meaning and context of Albert’s past feedback, and reached near-human accuracy in automatically tagging the category and sentiment for 100 % of incoming feedback. A more detailed, multi-label categorization was introduced.

Albert’s Results with Sentisquare

120 high-skill man-days saved yearly through automated classification. Time needed to deliver feedback report reduced from 3 weeks to 3 days. Based on the feedback coverage, each department, branch and region can laser-focus resources to make the greatest impact on customer satisfaction.

• 12,000+ open feedback pieces / month

• Multiple touchpoints and 3rd party services to monitor

• Subtle differences between topics – difficult to classify

E.ON’s Challenge

Open feedback consisting of unstructured text is crucial to understanding the customer’s view but difficult to classify automatically. For E.ON’s purposes, differences between key operational categories are subtle; a keyword-based text classification model would take years to build.

SentiSquare’s Solution

E.ON receives NPS and feedback from different sources at multiple touchpoints. The task was to bring all of them together to provide a 360° view. By learning to understand the meaning and context of E.ON’s past feedback, SentiSquare AI created a topic and sentiment recognition model with near-human accuracy.

E.ON’s Results with Sentisquare

With SentiSquare AI recognising the topic & sentiment of incoming feedback, E.ON can now measure the drivers of NPS at every touchpoint, down to each region and 3rd party contractor. Importantly, E.ON’s management makes better- informed decisions enabled by focused insight into all important indicators!

• 15,000+ feedback SMS / month

• Struggling to measure what drives customer satisfaction

• Biased & siloed customer intelligence

T‍-‍Mobile’s Challenge

Despite the ample feedback, managers struggled to measure causes of customer satisfaction — they lacked a quantified view; knowledge was biased and siloed within teams. T‍-‍Mobile tried out multiple state-of-the-art rules-based text analytics solutions. However, none could deliver a reliable output.

SentiSquare’s Solution

To provide a clear overview, SentiSquare deployed machine learning on T‍-‍Mobile’s data, resulting in high precision in topic & sentiment recognition. An analytics tool makes the output easy to visualise, interpret, and export. It also enables search for specific examples of customer feedback.

T‍-‍Mobile’s Results with Sentisquare

With SentiSquare AI now recognising the topic and sentiment of incoming feedback, T‍-‍Mobile can measure the drivers of customer satisfaction in near-real time. Importantly, T‍-‍Mobile’s management makes better-informed decisions based on focused insight on any important issue available any time.