Choose which optional cookies and similar storage we may use. Strictly necessary storage is always on, because the site cannot work without it.

Built into Averra

Prime Edge AI: intelligence that knows where to think

Prime Edge AI is Gateway Global's AI routing technology. A router on the device reads each question and decides where it should be answered: on the phone in your hand, or by a frontier model in the cloud when the question needs it.

Not every health question is the same size. Prime Edge AI decides where each one is answered.

Last updated

Answered on the device, or sent to the cloud only when needed.

How it works

Choose a request to see where Prime Edge AI sends it.

Diagram: a question is read by a router on the device. Routine requests, such as daily check-ins, meal logs and routine readings, are answered on the device. Harder questions, such as years of history, clinical literature or a live search, go to a frontier model in the cloud, only when needed. The person sees one conversation. On the device In the cloud A question Router on the device reads it first Answered on the device offline, no cost per call Frontier model only when needed The person sees one conversation.
  1. Reads the question on the device

    A router on the device reads each question before anything leaves the device.

  2. Stays on the device for routine tasks

    Daily check-ins, meal logs and routine readings are answered by Apple's on-device model, through its Foundation Models framework: offline, and with no cost per call.

  3. Goes to the cloud only for harder questions

    Questions that need years of history, clinical literature or a live search go to a frontier model in the cloud, only when needed. The person sees one conversation.

Trust built in

  • Every device attests itself in hardware.
  • No key ships in the app.
  • Access tokens expire hourly and carry no identity.

Why it matters

Privacy by architecture
Routine processing never leaves the handset. When a question needs the cloud, it goes straight to the model with no intermediary.
Economics that scale
Requests answered on the device have no cloud cost, so spend tracks the difficulty of questions, not the number of people asking them.

Apple first, Android next

Draft (D4): Softened wording on why Apple first.

Apple silicon offers fast on-device processing, and one Foundation Models framework spans iPhone, iPad, Mac and Watch.

Android support, using Google's on-device foundation models, is planned next.

Where it runs

Prime Edge AI is built into Averra, the digital health twin we are developing.

Read about Averra
Draft (D4): Whether Prime Edge AI is offered for clients' products.

[TO CONFIRM: availability for partners' and clients' products]

The one-page technology brief

The Prime Edge AI brief sets out the design on a single page, and is published here in full as HTML.

Read the brief
Draft (F7): Download the brief (PDF, 1 page), once the PDF is in the repository.

Download the brief (PDF, 1 page)

Questions people ask

What is model routing?

Model routing means choosing, for each request, which AI model should answer it. Prime Edge AI reads each question on the device and decides whether it can be answered there or needs a frontier model in the cloud.

What does on-device AI mean?

On-device AI runs a model on the phone, tablet or computer itself rather than on a server, so a routine request can be answered without leaving the device.

Draft (F7): The exact answer to "Does health data leave the phone?"
Does health data leave the phone?

[TO CONFIRM: The exact answer to "Does health data leave the phone?"]

What is hardware device attestation?

Hardware device attestation lets a device prove, using keys held in its secure hardware, that it is a genuine device running a genuine copy of the app. With Prime Edge AI every device attests itself in hardware, so no key needs to ship in the app.

Draft (D4): Cloud model provider, named only if approved.
Which model answers in the cloud?

[TO CONFIRM: Cloud model provider, named only if approved.]

Draft (D4): Whether Prime Edge AI is available for partners' and clients' products.
Can we use Prime Edge AI in our app?

[TO CONFIRM: Whether Prime Edge AI is available for partners' and clients' products.]

Build with the people who designed it.