Conversational Interfaces @ BASF R&D

Image Description

Kirsten Prill

Image Description

Yusuf Emre Guenes

Image Description

Christina Speck

Description

Apps on a mobil device or on a computer are just one way of interacting with 'scientific functionality' (information retrieval, data entry, calculations,...).
The challenge is to develop novel/ conversational ways in which R&D colleagues can interact with the scientific functionality in the BASF App Store.

Design Canvas

Justificationary Knowledge

Conversational interfaces such as text-based chatbots or interfaces with voice recognition are widely used in service applications. Even if the number of conversational interfaces is increasing not all of them are a success. The use of chatbots has to be carefully though trought in order to ensure a good user experience.

Problem

The navigation process in the App Store is not clearly defined. No guidance is provided to support the user. This leads to a low user-satisfaction and creates a barrier between users and the App Store.

Objective

The objective is to make searching for Apps on the App Store more conversational, intuitive and appealing for users and to improve the user experience. At any time, the user should be encouraged to interact with the App Store. In addition to that, the interaction with the App Store should create satisfaction and support the user in his/her work.

Idea

To achieve this goal, we plan to design a chatbot interface for the App Store that enables the user to interact with the App Store via text chatbot or voice recognition. To guarantee a smooth interaction with the chat bot a navigation process for the App search is defined and implemented.

Context

A company in the chemistry sector is establishing an App Store in the research and development department to give R&D people access to Apps that are specifically developed to support them in their work. The web-based Apps can have functionalities such as calculation, research support or support in conducting and analyzing lab experiments.

Scope
Design Process

To perform the design task the Double Diamond method is used. The method is divided into four phases to structure the process: Discover, Define, Develop and Deliver. The problem space is opened in the Discover phase. In the Define phase the task is narrowed down to one specific problem. Many ideas and solutions for this problem are collected in the Develop phase. In the end one specific solutions should be concretized in the Deliver phase and implemented as a result of the design process. The approach is strictly user-centered and aims to analyze and fulfill the needs of the potential customer.

Evaluation

A Cognitive Walktrough was conducted in order to evaluate whether the chatbot is appealing and intuitive for potential users. It is important to evaluate this under the assumption that users are unexperienced and do not receive a special training to ensure that the tool is self-explaining.

Method
Artifact

A mock-up of the chatbot and the conversation flow was designed with the help of botmock.com. A tool to design conversation flows and mock-up videos with a drag-and-drop editor. In addition to that, a prototype was coded in Javascript while using Windows services to test the functionality of speech recognition.

Result
Design Knowledge

In retrospect, several feedback loops are necessary to develop a chatbot interface that has a good user experience and can also be implemented in a real-world software environment. A major challenge is to design an interface that is intuitive and attractive for unexperienced user.

Iteration 1
Empathize
Planning the project

Getting a grip of the assignment and figuring out what kind of tools to use and how to arrange our team work.

Discovering the App Store

Getting to know the App Store and its usage through interviewing the developers of the App Store, visiting BASF and conducting a survey.

Developing the Problem Space

Finding problems that users may have during the usage of the App Store, based on the our interviews with developers. Then gathering some solutions that may address our discovered problems.

Define
Problem Specification

Based on the comments on our first presentation, we conduct literature research to decide on the problem we want to address and its proper phrasing.

Preliminary Solution

Listing all the advantages our preliminary solution will have on the problem, drawn upon our literature research.

Ideate
Expert Interview

Interviewing an expert in the field of chatbot design about the options that we have with chatbot design.

Specifiying Chatbot Characteristics

Specifying characteristics of the chatbot according to the findings from literature review, expert interview, and the field visit.

Prototype
Creating mock-up

Creating a chatbot mock-up with the software of botmock.com. Implementing characteristics that were specified before.

Test implementation of mock-up

Implementing chatbot as a program in the App Store environment. Testing the functionality of speech recognition.

Test
User tests

Ask users to try out prototype and give feedback.

© 2024 MyResearchProcess. All Rights Reserved.