Public reference cases we use to anchor advisory conversations. No invented client logos. No imagined testimonials.
Programmes that show what credible AI-enabled service design looks like.
A nationwide data exchange layer connecting public and private systems through verified digital identity.
Interoperability and digital identity can make services more coherent before advanced AI is added — and AI is far more useful on top of coherent infrastructure than fragmented systems.
Reference architecture for public-sector clients exploring foundations for AI-enabled services.
National programme combining digital public infrastructure, service redesign, and central coordination.
Central coordination and digital public infrastructure can support AI-enabled public services — but require explicit safeguards on data, consent, and contestability.
Useful reference for government clients designing AI-enabled service delivery at scale.
Life-event-based service design that organises support around citizen journeys rather than departmental portals.
Shifting from departmental portals to human journeys changes what AI is for — from search assistant to journey orchestrator.
Anchors service-design conversations with public-service and NGO clients.
National AI capacity treated as compute, datasets, models, skills, startup finance, applications, and safe/trusted AI.
National AI capacity is a stack, not a single decision — and each layer needs its own governance.
Useful framing for ministries, regulators, and development institutions.
Patterns to recognise — and design against — before they take root.
Automated welfare debt inference deployed without lawful, fair, human-centred process.
Automation of consequential decisions without due process, contestability, or human accountability for harm.
Used as a cautionary reference when designing decision rights and non-automation registers.
Risk-scoring and institutional suspicion harmed thousands of families and amplified discrimination.
Risk models combined with institutional bias and weak contestability cause systemic harm.
Reference for algorithmic impact assessments and discrimination audits.
Predictive systems that direct enforcement based on historical data already shaped by enforcement choices.
Feedback loops that send enforcement to already over-policed communities and entrench existing patterns.
Reference for evaluating any system whose outputs change the inputs it later receives.
Sensitive data, hallucinated analysis, and inconsistent review spreading through informal AI adoption.
Tooling deployed without decision rights, evaluation, or audit trails — creating quiet operational and legal exposure.
Common starting point for enterprise governance engagements.
Made with Emergent