RESEARCHProducts built around durable boundaries.
Start with a bounded scope.
These are proposed starting ranges for scoping a pilot or integration, not checkout prices or commitments to deliver an unvalidated product. Final scope, taxes and commercial terms are agreed in writing. Hardware, data licences, third-party services and verified-value terms are quoted separately.
VALO30 retains its Norway-specific target of NOK 1,000,000 including VAT. Site, Foundation and infrastructure are separate. See Habitat scope → Discuss a pilot →
Use cases
Monitor regulatory changes, supplier signals or market developments through a bounded intelligence mission.
Intended impact · to validate
Less repeated searching and earlier awareness of relevant changes. Measure signal relevance, missed events and analyst time in the pilot.
Use cases
Govern payments, access changes and physical commands proposed by people or AI agents.
Intended impact · to validate
Prevent actions outside current authority and make allow, deny or escalation decisions inspectable. Validate each action boundary and integration.
Use cases
Map invoice handling, case preparation or service workflows before selecting automation.
Intended impact · to validate
Identify ownership gaps, exceptions and suitable changes. Evaluate rework, cycle time and exception handling against the existing workflow.
Use cases
Enforce exact authorized effects at API, tool or device boundaries; handle revocation and stale approval.
Intended impact · to validate
Bind permission to the consequence that actually occurs. Test bypass resistance, replay and revocation behavior for the integration.
Use cases
Continue research, software work or case analysis across sessions and model providers.
Intended impact · to validate
Reduce repeated setup and loss of working context. Measure recovery time and continuity errors in representative work.
Use cases
Explore personal assistance that retains relationships and working preferences across providers.
Intended impact · to validate
More consistent assistance as models change. This is a concept; continuity and user control require validation.
Use cases
Compare demonstrated digital-worker capabilities and select candidates for bounded assignments.
Intended impact · to validate
Make capability matching explicit and reduce unsuitable assignments. Matching quality and task outcomes remain evaluation targets.
Use cases
Source human, digital or hybrid capacity for a defined assignment with acceptance criteria.
Intended impact · to validate
Make delivery and settlement depend on verified completion. Evaluate completion quality, disputes and total cost per assignment.
Use cases
Explore shared care routines, activity records and selective information sharing for one dog.
Intended impact · to validate
Help owners coordinate care and notice changes. Benefits require user studies and veterinary review where health claims are involved.
Use cases
Assess modular housing, shared facilities and relocation options for a specific site.
Intended impact · to validate
Improve flexibility as local demand changes. Site feasibility, full cost, permitting and occupancy must be assessed before claiming value.
Use cases
Prototype company formation and continuity of company identity across regulated providers.
Intended impact · to validate
Reduce repeated administration and fragmented records. Validate provider integration and jurisdiction-specific requirements.
Use cases
Keep a principal’s identity and address consistent as devices, accounts and communication channels change.
Intended impact · to validate
Reduce fragmentation and repeated setup. Portability, recovery and permission handling must be validated.
Use cases
Evaluate a real AI workload against a locked quality, cost, latency or throughput baseline.
Intended impact · to validate
Retain only verified improvements within agreed constraints. Savings are case-specific; no general percentage is claimed.
Use cases
Explore repeatable delivery of document preparation, reconciliation or other bounded operational work.
Intended impact · to validate
Make output, constraints and completion evidence comparable across runs. Measure quality, exceptions and cost per accepted outcome.
Use cases
Test local inference and controlled physical effects on constrained target hardware.
Intended impact · to validate
Potentially reduce latency and dependence on remote compute. Hardware performance, reliability and feedback binding require target-specific tests.
Use cases
Explore shared-room assistance, interpretation and translation using a common room context.
Intended impact · to validate
Potentially reduce repeated explanation and context mismatch. The prototype’s connected AI behavior must be demonstrated before benefits are claimed.
Use cases
Follow continuing stories, music and episodes around Olav and his family.
Intended impact · to validate
Build continuity across creative experiences. Evaluate audience engagement and return visits; no audience outcome is promised.
Use cases
Turn source material into traceable lessons, exercises and learning content for educator review.
Intended impact · to validate
Support repeatable content production. Measure educator effort, content accuracy and learning outcomes separately.
Different responsibilities address different failures.
A recommendation can be useful without being authorized. A tool can be connected without having permission for the proposed effect. An authorized command can execute without proving the intended result. CAI, AI Bridge, HEIMEL/REHT and evidence keep those decisions separate.
The same separation explains the wider map: identity survives provider changes; continuity preserves work across sessions; EXUP optimizes within an accepted outcome; GCU settlement requires supported completion.
Read why each layer existsGCU · Governed Completion Unit
One bounded piece of work complete under its Governed Contract: evidence present, constraints satisfied, authority valid and outcome recorded.