Exploration
Why isn't the messenger being used? Where could it be simplified?
- Conducted interviews and retrospectives to derive actions and improvements.
- Analysed the ticket data to find patterns.
With conducting interviews, retrospectives and analysing the ticket data, I tried to find indicators of how the chatbot might be improvable. In the beginning the question was also to even keep it at all. Next to the user perspective it was therefore also crucial to understand how it could provide value and efficiency to the customer service team.

Definition
Which decisions and assumptions did I follow up with?
- Boil down the scope and functionality of the chatbot.
- Setting the focus to the key interactions and touchpoints.
One of the main insights was, that probably typical for a corporate product, the 1892messenger had a lot of features both facing the user and the customer service team. It was for example integrated in their SAP environment to at some point automatically create actions within SAP. BUT most of that functionality was either never fully implemented, never really used or worst case, leading to a lesser usage. The main decision was to actually reduce the chatbot functionality.
Development
What did I end up defining?
- Improved the user flow regarding most-common requests and information.
- Improved the integration to cater the common workflows of the customer service team.
- Prepared a small campaign to push the new 1892messenger.
The easy part was cutting the functionality towards SAP and replacing it with an integration for Outlook, where (unfortunately) the customer service team worked on a daily basis. For the user I reviewed the flow for typical ticket types and information requests while setting the focus to the key information the customer service team needs to quickly tackle a ticket coming from the chatbot. Next to the functional improvements it became clear that the new chatbot should again be pushed with a small campaign and to expand its usage as a chatbot on the website as well.



Learning
What did I take away? What will I do differently next time?
In all the projects related to digital transformation or previously broader Quality Management, I personally struggled with properly finishing a project and going all the way from A to Z. With this project I finally had the opportunity to actually go from process auditing to improvement exploration to improved product, which was very satisfying. One of the key difference must have been to heavily focus on the one process and not drift off to the 10 related processes and reducing the improvement to the most valuable starting point. Onviously the chatbot wasn’t afterwards perfected but it created a new value and had a more solid foundation to move forward.