From: aidotengineer

Alma, an AI nutrition companion, has gained significant insights into improving user experience with multimodal interaction through its development process [00:07:25]. Initially, there was a focus on determining a single “winning” modality for user interaction [00:07:07].

Benefits of Multimodal Interaction

The development team at Alma discovered that users highly value the flexibility of multimodal AI systems [00:07:29]. Instead of prioritizing one specific modality, it’s crucial to offer users a variety of options that make sense for their context [00:07:40].

Examples of Modalities in Alma

Alma enables users to interact in several ways for nutrition tracking:

  • Voice The voice interface allows users to track meals quickly, such as logging breakfast, lunch, and dinner in under 10 seconds, transforming a previously laborious exercise into a simple one [00:07:09].
  • Photo Users can take a picture of their food [00:07:37].
  • Text Users can text their food entries to Alma [00:07:39].

Users appreciate that Alma makes it easy to “talk to, to take a photo, or to text,” adapting to their current situation [00:07:35]. This flexibility contributes significantly to a positive user experience [00:07:25].

NOTE

The key lesson learned is to provide users with as many different modalities as are relevant and sensible for their needs [00:07:44].