Self Service Customer Support of Electric Vehicle Charging Stations

Sammanfattning: The aim of this master thesis is to develop a suggested methodology for how to use Mavenoid infrastructure to improve customer support of DEFA EV chargers. Mavenoid is a company that helps other companies automate customer support, especially troubleshooting. This is done with Mavenoid models, interactive selfhelp tools that guide end users without technical knowledge through the troubleshooting process. Mavenoid models provide value both by deflecting cases (the end user solves the problem on their own using the model) and triaging cases (collect relevant information about the problem before escalating the case to a human support agent) The main methodology to develop a suggested methodology was learning by doing, using the suggested methodology to actually implement Mavenoid models available to end users on DEFA’s home page. This was complemented with a literature review, interviews and data analysis from model usage. The suggested methodology is to iteratively follow the steps of deciding which models to build, make priorities within these models, build the models, analyze their performance and continuously improve the models. To decide models, carefully evaluate DEFA’s support situation to decide where Mavenoid models would have the greatest impact. Force yourself to make quantitative assumptions to estimate a payback time for each possible model. For each model, carefully prioritize what to include and where the focus should be using estimates of frequency, value and time to model. Build the models to maximize deflection and triage and minimize abandoned sessions. Collect and analyze data from model usage and use this information to improve the models. To prioritize between possible improvements, force yourself to make quantitative assumptions of value and time to model and rank improvements by payback time. Limit the improvements you make either by time available or desired payback time. The potential business opportunity between Mavenoid and its customers is more attractive the more support cases the customer has and the larger fraction of end users that use Mavenoid. The business opportunity varies greatly with assumptions that are very difficult to estimate accurately at the early stages of a Mavenoid implementation. This indicates that Mavenoid models should be implemented step by step and assumptions updated when more data is available. Implementing Mavenoid models can be both positive and negative from a sustainable development perspective. They could encourage people to repair products instead of replacing them, scale renewable energy technology faster and remove boring and repetitive tasks from support staff. On the other hand, they might not be appreciated by all end users, could lead to increased electricity consumption and potential unemployment for support staff. Being about a largely unresearched topic, the results in this thesis are relatively subjective. This suggested methodology was used and proved to work to implement Mavenoid models for DEFA EV charging stations but it should be seen as one possible methodology, not the confirmed best methodology.