Goethe-Institut USA · Education & Culture
Chatbot or a better website? Research for the Goethe-Institut USA.
- Period
- 01/2021 – 06/2021
- My role
- UX lead
- Team & context
- A junior UX designer, the Goethe-Institut USA site manager and a contact at UserTesting.
Challenge & approach
For the Goethe-Institut USA, I evaluated whether a planned chatbot would help prospective students find the right German course. The goal was to clarify, before costly development, whether the target group would accept a chatbot at all.
Finding the right course on the website took a lot of effort, and staff had to call prospective students to assess their language level and availability. A chatbot was meant to take this over, but there were doubts that the target group would accept it. Staff were skeptical too, because prospective students particularly valued personal one-on-one advice. So: validate before building.
Research & methods
14 remote in-depth interviews of 60 minutes each with 8 Goethe-Institut staff members, 3 teachers and 3 learners in the US laid the foundation. They led to a to-be user journey, a concept in Balsamiq and a fully interactive chatbot prototype in Axure RP, embedded in the real website.
The unmoderated UserTesting study covered 8 participants from Germany and Canada: 4 on desktop and 4 on smartphone. After 2 pilot tests, we adjusted the study setup. I analyzed the sessions and edited the most telling scenes into highlight videos for the stakeholders.
Result
Solving the right problem.
The target group did not want a chatbot. The real problem was the website, which completely overwhelmed users. With the highlight videos, I convinced the stakeholders to drop the chatbot development; the budget saved went into improving the website instead.
I would take the same approach with AI chatbots today: understand the problem first, then choose the technology.


