Yellowtail organized its annual business relationship day on September 21, 2018. An important theme during this day was 'Robotising'. Especially the design and development of a chatbot. Together with VGZ, our robo team gave an instructive workshop consisting of:
1) How to build a chatbot (technique),
2) How to design a conversation (Finversation design)
3) What does a good working chatbot look like in practice (VGZ - SAM).
The case study was presented by VGZ and consisted of a behind-the-scenes peek at their self-built chatbot, SAM. During this day, Yellowtail and VGZ discovered that they have a lot in common in the field of robotization but can also benefit from sharing their knowledge. For example, VGZ first invited Yellowtail to transfer knowledge about conversational design. In order to be able to transfer this knowledge properly, Yellowtail has developed 3 workshops, all with their own theme on conversational design.
With chatbot SAM, VGZ has a considerable head start on innovation in the insurance market. After a successful go-live and positive results, it was time to continue developing. Where they had always designed conversations for a large screen, they now have to take other things into account, for example, the emerging voice. The challenge: designing and building conversations for the various conversational devices: voice, at a glance (e.g. google home hub or in the car) and full screen. To help VGZ with this, Yellowtail has developed three workshops based on the Finversation Design theory. A unique combination of Conversation Design and Behavioural Economics.
Together with VGZ, we have developed sample interviews for a complex case within the health insurance sector. We developed these interviews for voice, at a glance and full screen. VGZ will use these developed conversations and the learned methodologies to expand their chat assistant for voice even further - and to develop them for the Google Home, for example.
Head of Experience Design
Stefan Visser - Digital Cooperation Catalyst “Through these workshops and refreshing assignments during the workshops, we had the ability to look differently at the difficult matter of funding for health insurance, namely more by theory and psychology. These insights will be used straight away to try if we can setup a new and improved chat engine, which can be used directly for several channels.
Like mentioned before, the delivered product consisted of three workshops, each having their own theme.
The human conversation In this workshop we first had a look at what the Finversation Design theory contains, subsequently, with this theory in the back of our minds, we had a look at how the human conversation actually goes. We did all of this to be able to firstly set the preconditions for a good conversation (case, user persona, system persona, etc.)
After the theory part, we also executed a practical assignment. Namely acting the case, the user persona and the system persona. Where one person played the user persona and the other the system persona. These conversations have then been recorded by video. We did this so we could experience what is relevant in a chat conversation or not.
Test & improve The conversations from workshop 1 have been processed, subsequently we had a look at all comparisons throughout them. The most common conversations were summarized and used as the ground work for the chat conversation. Thereafter we had a look at the ‘Nuding Theory’ and how we can apply this theory to the summarized conversations. Lastly we did a Wizard of Oz test with users by phone.
Upscale: In this last workshop we focussed on the use of conversations for different devices. For this we applied ‘Design Scaling’.