What would the perfect AI for natural language customer experience be like?

  • Super accurate,
  • when confused it asks the right clarifying question,
  • easy for anyone to teach, design, edit, and control,
  • speaks all languages,
  • and immediately learns new skills/domains.


Gamalon Architecture

Idea Studio Workflow

  1. The system ingests your raw customer messages, and proposes an Idea Tree.  (Or, if you like, set up a desired target schema yourself)
  2. Collaborate with the AI and with your colleagues on the Idea Tree – it’s like Google Docs for AI model creation.
  3. As new messages come in, drill down multiple levels to very specific categories and features or just monitor at a higher level.
  4. Watch it flag messages that are not well covered by the current Idea Tree to spot trends, mistakes, and critical errors before it’s too late.
  5. Collaborate with the AI on the Idea Tree so that it can correctly process these new kinds of messages.
  6. Goto 3