Technology & partners

Model-neutral by design, not by accident

We are not committed to a single provider. Each component of an agent uses the model that performs best for that task — Arabic comprehension, document processing, voice, or reasoning — and can be replaced as the field advances, without rebuilding the agent around it.

Last updated: 12 August 2026
01

Why neutrality matters

The provider leading today may not lead in six months, and may retire a model you depend on. When an agent is built around one provider, any change on their side becomes a rebuild on yours. Because we select per task, replacing a component stays a configuration change rather than an architectural one.

  • Model selected per task: Arabic comprehension, documents, voice, reasoning
  • Any component replaceable without rebuilding the agent around it
  • No lock-in to a single provider contract
02

Integration surfaces

The agent operates inside the systems a business already runs.

  • Websites and mobile applications, via script tag or SDK
  • E-commerce platforms: Shopify, WooCommerce, Salla, Zid
  • Booking, clinic and point-of-sale systems
  • Google and Microsoft calendars
  • WhatsApp Business, web chat, telephony and email
  • CRM platforms, read and write, with field-level rules
  • ERP, billing and accounting systems via API
03

Governance & security

Access is granted at the least privilege sufficient to perform the task, and every action an agent takes is written to a complete audit log.

  • Data boundaries and retention agreed per engagement
  • Least-privilege access to client systems
  • A complete audit log of every action an agent takes
  • Published Responsible AI and Trust & Security policies