Kärcher drives customer engagement with Contentful and Alexa Skills

Company Size

3000

Year Founded

1935

Headquarters

Winnenden, Germany

2

new Alexa skills launched

faster to iterate and edit content

3

languages localized in Alexa Skills
Plan Type
Team
Share Case Study
Main Challenges
  • Content storage separated from distribution with easy access for editors

  • Flexible content types for unconventional distribution channels

  • Ability to update live content in several languages within a published Alexa Skill

Solutions
  • Unique user experience for customers powered by Contentful and Amazon Alexa

  • Customers can access information and interact with Kärcher products via voice control

The integration of Contentful reduced the complexity of our architecture due to the excellent API and technical documentation in combination with the convenient backend for data modeling.

Jens Hildenbrand, Program Portfolio Manager

Kärcher

Project Story

Building Alexa Skills with Contentful

Content models need to meet the standards of the voice provider and remain accessible to editors. If the models are not built correctly, Alexa cannot process prompts nor return the requested information. Editors need to publish content in several languages and have long-term, easy access to it, especially if updates are needed.

Zoi handled the market research with a cross-functional team. After a series of reviews, they chose Contentful for its flexibility, security and user-friendly interface to solve the content challenge. The team began by building a proof of concept before moving on to creating a content model for the full-scale project.

After building the content model, Zoi developers were able to work simultaneously with Kärcher editors. The developers connected the GraphQL Content API and the Alexa API, which let editors manage voice content in the custom content model through the Contentful web app. Editors began rewriting and condensing content. For voice assistance, brief statements are preferable (it's far easier to read long sentences than listen to them).

With their content stored in Contentful, iterating was painless. The team ran through three additional iterations of the content model with user testing. Working this way enabled them to continually make any necessary adjustments. They learned helpful, unexpected information this way. For example, slot values can only be used once. Using the same word multiple times caused the Alexa to stall.

Ask Alexa for Kärcher cleaning tips and control devices via voice control

With Contentful, Kärcher developed and released two Alexa Skills. Want to control the irrigation in your garden or find the right high-pressure washer? It’s simple. The first Kärcher Alexa Skill allows you to control the Watering System Duo Smart Kit with Alexa. And, discovering the perfect Kärcher product or cleaning advice has never been easier with the Kärcher Info Skill.

Ease of localization in Contentful makes scaling across markets and languages simple. The Control Device Skill is now available in four markets and three languages: German, English and French. The Content Skill is also available in several markets and in two languages: German and English. Their content can now be reused chatbots and the Kärcher IoT-enabled machines.

"The integration of Contentful reduced the complexity of our architecture due to the excellent API and technical documentation in combination with the convenient backend for data modeling."
- Jens Hildenbrand, Program Portfolio Manager , Kärcher

Key learnings

  • Test repetitively with real people. Include feedback from users testing the product from the beginning of development. This process ensures your product remains user friendly and solves actual challenges.

  • Voice systems are built on different content structures than traditional apps and websites. They can also be used in chatbots and help pages. Limiting the content length is critical.

  • Minimize complexity in your content structure to reduce barriers between editors and content.

add-circle arrow-right remove style-two-pin-marker subtract-circle