Mobile Considerations - Chat-based Assistants

The fifth post in a series on the developments taking place within mobile technology.


In recent years, more and more organisations are moving towards messaging-based automated assistants. This may be partly due to the aforementioned app fatigue but also shows an increased understanding of consumers’ digital habits; organisations are finally meeting the customer where they already are.

What are Chat-based Assistants?

They are automated messaging systems that imitate human interaction.

The dominant social media platforms all allow organisations to directly message potential or existing customers. More recently these platforms have built tools that allow customers to interact with autonomous, chat-based, digital assistants.

The main platform, Facebook Messenger, provides a large suite of tools. Extremely rich chat-based experiences can be created; customers can browse products, pay for items and receive tickets/passes.

The chat-based assistant platforms are continually expanding to enable even more engaging experiences but many of the underlying development tools are multi- platform.

Additionally, components of a chat-based assistant can be reused across the multitude of voice assistant platforms, such as Amazon’s Alexa.

The most effective chat-based assistants are focused on specific customer outcomes (e.g. reserving parking). Extremely simple assistants can be created using existing content, such as FAQs, and then later expanded to serve common customer needs.

As these assistants are typically deployed within social media platforms, they are multi-platform (including mobile and desktop) from launch.

Benefits of Chat-based Assistants

  • Direct, one-to-one customer relationship
  • Can leverage device functionality, such as cameras, through the native chat apps
  • Low level of friction to engage users

Delivery Challenges

  • Unconstrained user input requires the experience to fail gracefully when the assistant can’t handle the request
  • Chats are easily ignored/deleted
  • Platforms are still maturing

Case Study – KLM

Chat-based assistants are well-suited to supporting or automating basic customer service queries and can be trained using existing structured data (e.g. FAQs) or by observing the actions of a customer service agent.

Starting with flight details and more recently expanding to ticket sales and local directions, KLM’s Facebook Messenger bot leverages the wide install base of the Messenger app to deliver an innovative and personalised customer experience.

The Messenger-based bot was introduced to support KLM’s social media customer support teams, who handled a significant volume of incoming customer messages via Facebook andTwitter. The bot was designed to learn based on customer service agents’ actions and slowly the responses were automated.

The bot’s features were expanded beyond customer service queries to allow automated flight information alerts to be delivered and the retrieval of boarding passes.

User Benefits

  • KLM reported that 15% of online boarding passes are now sent via Messenger.
  • Additionally, the Messenger bot received a 5-point higher average Net Promoter Score (NPS) than other social platforms.
Written on August 6, 2018