Tailor-Made offers made-to-measure clothing and accessories. In order to provide such items, the application proposes tailor’s shop appointments via a conversational interface (chatbot).
This case study is part of a LINC research initiative focusing on interface design. Cases dealt with are fictional services co-created with the participants of the Data & Design workshops. The solutions considered here are not intended to be recommendations to copy, but rather contextual illustrations of creative processes that might inspire your services and products. The study does not cover the entire user experience, but simply concentrates on key points. As such, it does not necessarily cover all GDPR requirements.
Product’s Context
The Tailor-Made brand specialises in sale of made-to-measure garments. Its positioning assumes that customers go to one of its tailor’s shops to have their measurements taken. The mobile application enables users to choose one or more articles, make an appointment in the nearest shop and be notified once the product is ready. These interactions are guided by a conversational interface (also known as a chatbot). When the order is ready, it is delivered to the customer’s home.
The chatbot choice breaks from the interfaces ordinarily used in order to collect web users’ information. Getting away from form codes and boxes to tick enables provision of new purchasing and consent experiences. In this new conversational world, what would be the best architecture in compliance with users’ rights?
User pathway and key steps
The purchase pathway is unique in that it is divided into three stages. The first happens online, when users fill their shopping carts and make a tailor’s shop appointment. The appointment itself is the second stage: once at the shop, users’ measurements are taken and their orders invoiced. Finally, following the appointment, the service regularly notifies customers on the progress of their orders up until they are delivered. It is evident that most information is upwards in the first half of the pathway, when users are invited to provide information, and downwards in the second half, with notifications on the service’s monitoring of the order. The Tailor-Made team largely focused on the first half.
This example provided participants with an opportunity to take a closer look at forms of data collection and of collecting consent in the context of a conversational interface. They put forward several solutions, including:
- spreading collection of personal information across the conversation guided by the chatbot. The experience provided by use of a conversational interface leads users to share information useful to the service by avoiding giving them the off-putting impression of having to fill out a form. Such splitting up of information collection enables each step in collection to be combined with a request for consent;
- adding a confirmation step when payment is made, at a time in which users’ vigilance is usually increased as they have to carry out an important action that also commits them. With chatbots, where obtainment of consent is incorporated into the conversation, this stage makes it possible to check whether or not users have expressed their choices correctly.
The conversational interface: another category for collecting information
Participants saw the chatbot as an opportunity to contextualise collection of information by splitting the process up, as the choice of a conversational mode provided more occasions to inform and elicit users without affecting the quality of their experience with the service. Hence, discussion alternates between questions on users’ needs and wants (which item, what colour, what date for the appointment?), collection of information (name, location, email address), and details to do with data processing and requests for consent (receiving the newsletter, partners’ offers, geolocalised advertisements). In this way, participants found a fluid, contextual means of incorporating data collection and obtainment of consent into the user pathway.
Proposed approach
An extra step for checking how one’s data are shared
One of the participants wondered whether obtainment of consent via a conversational interface might represent a risk to users. If a user mistakenly chooses an item he/she does not want, the conversational interface does not immediately provide much means of changing their choice, unlike other kind of interactions (such as a preference management dashboard).
In order to deal with this type of situation, participants integrated a consent recap at the end of the appointment pathway. They chose to dissociate it from the conversation’s format and include it before final validation of the appointment request. By summarising the consents that were collected over the course of the conversation, the recap enables users to make sure that their information is shared in accordance with their preferences and to change a wrong choice made during the conversation with the chatbot. In the event of a user wishing to have more information on the way in which his/her data is used, a link to the confidentiality policy was also integrated at this point in the interface.
Proposed approach
Limitations
As participants in the workshop pointed out, the conversational format may be a double-edged sword for users. It may involve them more and increase the attention they pay to the service and their actions, but at the same time it may be highly distracting as it offers a new means of interaction more likely to produce operational errors on their part. With this new type of interface, it is essential to integrate ways of easily reconfiguring choices, included here at recap level. Nonetheless, conversational interfaces provide interesting possibilities for implementation of the GDPR’s principles, due to the need for sequencing and narration of information presented to users.