Model choice
A model-agnostic architecture designed to route work according to capability, cost, reliability, and data policy.
Sunrise Intelligence
Advanced intelligence. Simple on purpose.
Sunrise Intelligence is the developing technical infrastructure behind our product direction: connecting models, organizational context, tools, and human control so intelligence can participate usefully in real work.
The application layer of intelligence
The intelligence is only as useful as the system around it.
We are not developing a proprietary foundation model. Our work is in the architecture around models: durable context, tool access, orchestration, approval boundaries, and a clear record of what happened. The ambition is to make advanced intelligence accessible without asking every organization to build its own AI department.
Sunrise Intelligence / Conceptual architecture
Organizational dataExisting systemsPersistent context
Sunrise Intelligence
Context → Reasoning → Action
Models + agentsTools + workflows
Human approval · Permissions · Observability
SunriseGeoKnownWaterContinuum
Architecture / Development direction
The intelligence layer underneath the work.
Persistent context
Organizational knowledge and workflow state that can remain useful beyond a single conversation, with explicit access boundaries.
Agents & tools
Specialized agents connected to approved tools and integrations, with clearly bounded responsibilities.
Durable work
Workflow orchestration designed for tasks that need to pause, resume, recover, and wait for human approval.
Human control
Permissions, least privilege, review points, and escalation rules that make authority explicit.
Observable systems
Traceable actions, costs, exceptions, and outcomes so a useful system can also be evaluated and improved.
Development status
An honest view of where we are.
Deployed foundation
Operational systems
The internal Sunrise Geo deployment provides evidence for connected reporting and workflow improvements. It does not establish production readiness of the full Sunrise Intelligence platform.
Active development
Shared intelligence infrastructure
Our development direction brings model access, agents, context, integrations, and approval boundaries into a reusable foundation. It is not a generally available platform today.
Research / Future direction
Private & accelerated compute
We are exploring local inference, private and hybrid deployment, and future accelerated compute. Infrastructure investment should follow measured workload, privacy, reliability, and cost requirements.
Run intelligence where the work requires it.
Cloud models offer access to rapidly improving capabilities. Local or private compute may serve workloads with different privacy, latency, or economic needs. Our direction is to preserve that choice without coupling every workflow to one provider. Running locally does not remove the need for permissions, evaluation, or human accountability.
Read our responsible intelligence principlesApplied intelligence
The infrastructure matters when the work gets better.
Explore how this direction connects our AEC focus and the wider Sunrise portfolio.