Skip to content

Evidence before action

Automation is earned, not assumed. Every recommendation ships with provenance, every action is reversible, and every outcome is verified.

Governance philosophy: Assist → Auto

We believe automation must be earned through evidence, not assumed. Every deployment starts with full human oversight. Automation unlocks only where policy, confidence, and rollback conditions are already agreed.

Level 1

Assist

Humans approve every recommendation

  • Full human oversight of every decision
  • Platform provides recommendations with evidence
  • Teams review, approve, and execute
  • Feedback improves future recommendations
Level 2

Govern

Actions stay inside agreed policy bounds

  • Ranked action queues based on impact and confidence
  • Escalations and edge cases stay with people
  • Explanations and provenance are visible before action
  • Rollback conditions are defined in advance
Level 3Earned

Auto

Automation is earned for scoped actions

  • Autonomous execution only for approved action types
  • Exception-based human review
  • Continuous monitoring and anomaly detection
  • Rollback capability for any action

You control the pace. Every move toward automation requires explicit consent, clear policy bounds, and evidence that the action can be reversed if needed.

What every recommendation shows

Before any team acts, they can see what triggered the recommendation, what supports it, how confident the platform is, and what happens next. Nothing depends on hidden logic.

Drivers

What triggered this recommendation? Sensor data, weather forecast, tariff change, occupant schedule — every input is named.

Evidence

What data supports it? Raw readings, model outputs, and domain context are linked so nothing is taken on faith.

Confidence

How sure are we? A calibrated score with bounds, not a hidden probability. You see when the platform is uncertain.

Action

What should happen next? A concrete, reversible step — with rollback conditions defined before execution.

How this shows up in the products

Trust is not separate from the product experience. It is what lets teams act with confidence across design, delivery, and verification.

OEM Lifecycle Intelligence

Design assumptions, installation outputs, commissioning checks, and observed performance stay connected so OEM teams can see what changed and why.

Home Energy Intelligence

Programme teams can explain planning choices, support customers with evidence, and verify whether connected homes are performing as expected.

Building Estate Intelligence

Estate teams can prioritise interventions, coordinate work, and prove outcomes without losing the context behind each asset decision.

Data & IP boundaries

Clear boundaries protect your data and intellectual property. No surprises, no hidden terms.

Customer data stays with the customer

Raw data is processed in-situ or in isolated tenants. No training on customer data without explicit consent.

Evidence is portable

Customers can export their evidence packs, Model Assets, and audit trails at any time in human-readable and machine-readable formats.

No lock-in

Portable artefacts ensure continuity. Switch providers without losing your accumulated evidence and governance history.

IP boundaries are explicit

Customer data, customer-specific Model Assets, evidence packs, and configurations remain customer property. Platform software, core algorithms, and platform improvements remain Aeterno property.

Security & compliance

Security and compliance support need to hold up in procurement, delivery, and review. We design for that from the start.

SSO-ready

SAML 2.0 and OIDC support can be configured for enterprise identity providers

Role-based access

RBAC with viewpoint-specific permissions and audit trails

Audit trails

Action and access histories are recorded for review and export

Data Encryption

Encryption at rest and in transit using industry-standard protocols

Tenant isolation

Per-customer isolation with no data commingling between tenants

Regional deployment options

Deployment approach can be aligned to regional hosting and data-localisation requirements where needed

MCS

Evidence outputs can support commissioning and installation-quality reviews against MCS requirements

Building regulations

Evidence packs can support Part L and related compliance workflows

GDPR

Data processing agreements, privacy by design, and data subject rights

ISO 27001 controls

Security processes are designed to align with ISO 27001-style control expectations and customer due diligence

Frequently asked questions

Need to review governance, ownership, or deployment fit?

We can walk through how trust, evidence, and control work across your products, teams, and deployment requirements.

Talk to Aeterno