AN INNOVATIVE PROJECT AS THE BASIS FOR NEW BUSINESS MODELS

To develop new touchpoints for our client, we started with five sprints using the »Double Diamond« method, which we completed within nine months. The interdisciplinary team we used consisted of strategists, developers, analysts, and client team members from the digital marketing field. The topic of Voice User Interface was previously uncharted territory for our client.

THE VALUE OF REUSABLE CODE AND CONTENT

The project started with a comprehensive data analysis and an employee and customer survey, which provided insights into the questions potential buyers asked about our customer's product range. From the beginning, we tested with real people and immediately took their feedback into account. This helped us to create realistic and user-friendly content. 

The content we localized and structured during the design sprints, which we had collected and digitized from various sources such as catalogs, magazines, apps, websites, and brochures, was used for the client's other digital content areas in addition to the voice client. 

Right from the start, we also took care to ensure the voice systems require different content structures than conventional apps and websites. After a series of tests and feedback rounds, we selected the content management tool Contentful due to its flexibility, security, and user-friendly interface. We started with a Proof of Concept before finalizing the content model for the entire project. 

We combined various tools to build the voice user interface. These included Alexa Skill Kit, AWS Lambda, and Contentful. All components were connected with the appropriate APIs. The code architecture and the backend operated via Contentful were kept minimalistic for our customer to create reusable content structures. Thus, the content editors were able to manage and expand new content and publish it in multiple languages without extra assistance.

KEEP IT SIMPLE TO MAKE IT LAST

How do we find the right content and language to reach the target group? Firstly, we tried identifying ourselves as customers and used the online store and the customer hotline to find out about the product range. Learning by doing: We ordered the products that were to be connected to the VUI and tested them extensively. Every single step was documented and converted into hypotheses. We validated these hypotheses with about twenty employees from outside the product. This resulted in concrete user stories. From the user stories, functions for assembly instructions, product finder, information about the company, and entertainment functions could be derived.

The technical architecture of our solution.

The real challenge was to build a varied and quickly-reacting voice system. Along with it, we had to deal with the storage, management, and provisioning of the content required for the voice client. A scheme for the voice commands to which Alexa reacts had to be worked out first in order to maintain content lengths and avoid word duplication because duplicate words cannot be processed in one command. As a result of the project, the customer integrated the headless content management tool (HCMS) Contentful into the marketing. This integration also had a positive impact on other marketing projects.

VUI creates a new touchpoint that creates an individual user experience.

The structure of the content model is designed and built to be sustainable so that new integrations can be easily implemented in the future. This way, the customer experience can be further improved – whether through localized weather data, personal data such as calendar information and order histories, or through memberships in customer loyalty programs.

The functions were implemented on an agile production plan, where testing was carried out on a weekly basis. Before implementing the final content, the VUI was tested again with the employees so that any relevant feedback would flow into the final integration. 

The technical structure is as follows: Alexa communicates with the Alexa Voice Service when a voice command is received. This takes over the transformation of the speech into text (Speech-to-Text). The text is analyzed and assigned using the interaction model to call up the required functionality of the skill. The functionality is represented by the AWS Lambda service. Lambda has access to the APIs provided by Contentful. This then enables the appropriate content to be played.

VUI contributes to a unique user experience for end consumers, as they can access product advice and product information directly via voice control. In addition, the creation of a new touchpoint paves the way for further innovations.

Reusable content structures enable all participants to generate and implement new content independently.

Reusable content structures form the basis for a logical content metamodel that can be used independently – from the website to the magazine page.