Alphabet is a BMW Group Company leasing company cars, bicycles and vans. It offers mobility solutions and products for fleet management and fleet financing worldwide. They have a fleet of over 680.000 vehicles.
Drivers have a lot of questions, also outside working hours. To handle these questions Alphabet decided to set up a driver interaction center and develop a chatbot: Alphie.
Microsoft was responsible for the design, development, and implementation of the conversational bot infrastructure on Azure. Vectr.Consulting and Arinti worked together to support the team of Alphabet with a Data Translator.
Our goal? Craft the conversation in an intuitive and natural way for the user.
Developing the bot in 12 weeks is an intense job and demands a lot of engagement of the Alphabet team. That’s where the Data Translator comes in, serving as an innovation partner through the development of the bot.
“Conversation is key and it shouldn’t be an afterthought. Instead, before developing Alphie, we created a roadmap of the possible conversations users could have with the chatbot.” — Data translator.
Before Microsoft set up the architecture and started developing the bot, our data translator teamed up with Alphabet to:
- Set the bot framework and create the bot scope
- Distill Alphabets’ brand voice into the system persona (= the chatbot)
- Make conversational designs
The first step in creating this roadmap is setting up the framework during the ‘bot insight workshop’. During the workshop, we ask ourselves the following questions:
- Why are we doing this?
- What is our goal?
- What do we want to achieve?
- Who are we helping?
- Who is our audience?
- Which problems does this bot solves?
We learned that the Alphabet chatbot will become a platform where different departments meet.
The customer journey of the end-user
We interviewed employees of almost every business unit to learn more about the people of Alphabet, their daily struggles and most common workflows. We dived into real life customer communications, which enabled us to define both general and critical user journeys and capture the goals of users in given situations. Based on this understanding we identified the as-is and to-be customer journeys.
Functionalities and back-end integration
When drawing the first high-level mental models of the end users, we zoomed into the functionalities of the bot. We linked the customer journeys with technical processes and identified the technical complexity and features of the bot. This way we considered both the users’ needs and technological restraints.
This also gave us a clear insight into what should be in scope for a well-trained and mature bot.
Who is Alphie? Time to create a personality.
The chatbot represents Alphabet on a personal level and is part of the customer care, its personality is present in every aspect of the conversation.
We took time to shape this personality with employees of marketing, customer care, and some process analysts. Together we worked towards a rich personality and tone of voice that perfectly fits with the end-user and its context.
Defining this scope and personality is a crucial step before designing real conversations and deciding on word choice or actions in specific situations.
We worked closely together with a UX designer, an architect and a developer to create the conversational flows. Microsoft guided us through the possibilities of their Bot Platform, which was crucial in designing the flow. After all, each vendor has other building blocks, visual or conversational components.
A flow represents the organization of a conversation, based on the underlying logic of the customers’ needs. It is a detailed design that represents the complete user experience.
We want users to be successful when they’re talking to Alphie. Therefore we have to make sure that Alphie asks the right questions, in the right way, in the right moment and finally provides users with the right answer.
Today you can find our young Alphie in his new home, the Alphabet website
Thank you for reading!
This blog was originally posted on https://blog.arinti.be/conversational-design-chatbot-969d80261d5f