Structured email aims to allow email senders to add explicit machine-readable structured data to their messages. Based on that, email receivers and tools on their behalf can make more sense of email content and hence provide a more efficient and enjoyable email experience.
Most of recent AI technologies similarly try to automatically make sense of unstructured email content. Interpretation is hence not prescribed by senders, but inferred by AI algorithms. While this requires less effort and can be applied to any unstructured email, results may not always be fully accurate.
Essentially, approaches based on explicit metadata annotation (such as structured email) and AI technologies can complement each other well:
- Metadata can easily be used in high value use cases and in cases where original email content is available in structured form already. In addition, structured (meta)data can help improving the quality and accuracy of AI based tooling.
- Conversely, AI and machine learning approaches can contribute to the large-scale creation of structured data from semi-structured text.
First, data models for many typical use cases are already in place, such as by Schema.org.
Second, the decentral extensibility of structured data is a core feature. While this may lead to the fact that a structured email may contain structured data which a particular receiving email client does not understand, this is no different from the popular concept of “email attachments” (see MIME).