What We Do
Governance-first support for current, agentic, and frontier-facing AI systems in institutional and regulated environments.
AGCIH supports governments, regulators, oversight bodies, public-interest institutions, and selected regulated entities to move from digital and AI ambition to responsible and trusted implementation. Our work focuses on governance architecture, institutional readiness, risk oversight, and public trust so that technological deployment strengthens legitimacy, accountability, and long-term institutional continuity.
AGCIH works across three levels of AI governance. At the first level, we address current AI governance challenges already facing institutions, including procurement, data protection, legal practice, public-sector adoption, and readiness. At the second level, we focus on agentic AI governance, where systems do not merely advise but begin to act through workflows, tools, robotics, and delegated machine action. At the third level, we examine frontier-facing governance questions: whether public institutions can retain lawful authority, oversight, interruptibility, reconstructability, and accountability as AI systems become more capable, general-purpose, autonomous, and difficult to supervise.
What AGCIH Does
We develop governance frameworks, readiness diagnostics, guidance instruments, and structured capacity support that help institutions clarify mandates, decision authority, escalation pathways, risk controls, assurance mechanisms, and accountability continuity for high-impact systems. Where appropriate, AGCIH engages selectively with regulated entities when AI deployment intersects with systemic risk, rights protection, or public accountability.
What AGCIH Does Not Do
AGCIH does not design, deploy, procure, or regulate technology. We do not replace statutory mandates or perform regulatory functions. Our role is to strengthen governance infrastructure so institutions retain decision authority, supervisory clarity, and accountability when digital and AI-enabled systems are introduced.
Our Three Levels of AI Governance Work
From present-day implementation to agentic systems and frontier-facing public authority questions.
Level 1: Current AI Governance
This level focuses on AI and digital systems already entering institutions today. It includes current tools, procurement, data protection, legal practice, public-sector adoption, and institutional readiness.
- AI governance readiness assessments
- Procurement and vendor governance safeguards
- Data protection and lawful access controls
- Institutional oversight and accountability mapping
- Legal and public-sector AI governance support
Level 2: Agentic AI Governance
This level focuses on AI systems that do not merely advise, but act. These include AI agents, robotics, workflow automation, autonomous decision-support, tool-using systems, and delegated machine action.
- Delegated machine action governance
- Administrative override and escalation design
- Agentic workflow accountability safeguards
- Human decision authority and review pathways
- Risk controls for autonomous or semi-autonomous systems
Level 3: Frontier AI Governance
This level focuses on future-facing governance questions raised by AI systems that become more capable, general-purpose, autonomous, and difficult for institutions to supervise.
- Frontier-facing institutional readiness
- Public authority and advanced AI systems
- Sovereign administrative authority
- Institutional hosting capacity for advanced systems
- Governance of frontier AI entering state functions
Start Here
Public tools and foundational analysis for institutions exploring governance readiness.
Frontier AI Governance | Latest Article
Before AI Goes Rogue: Frontier AI Governance and the Institutional Duty to Retain Authority
AGCIH's flagship frontier AI governance article. Argues that capability acquired without governability is not modernisation — it is institutional abdication. Applies Administrative Hosting Capacity, Relocation of Judgment, Continuous Administration, and Sovereign Administrative Authority to frontier-facing contexts, with a specific focus on how advanced AI enters African public administration through procurement, platforms, and infrastructure dependency.
Download Public Tools
Access AGCIH public governance tools including the Child-Safe AI Toolkit Lite, FRIA-Lite, Public Trust Controls Pack, Agentic AI Readiness Checklist, and the Institutional Engagement Pack.
Read: Knowledge Paper 001
Rule of Law in the Age of Agentic AI explains why accountability must be structurally rebuilt before institutions automate authority. The paper introduces five rule-of-law tests: legality, accountability, reasons-giving, contestability, and auditability.
AGCIH publishes selected public resources while reserving detailed diagnostics, implementation instruments, and institutional templates for structured engagements.
AI Governance, Data and Public Authority Programme
Strengthening governance safeguards and accountability for current, agentic, and frontier-facing AI systems.
This programme supports institutions to establish governance arrangements and structured safeguards for AI and advanced data systems, including oversight pathways, escalation controls, institutional accountability mechanisms, and public authority safeguards.
The programme recognises that AI governance is no longer only about responsible use of current tools. Institutions must also prepare for systems that act through workflows, rely on external infrastructure, and may become more capable or difficult to supervise over time.
- Governance frameworks for high-impact digital and AI-enabled systems
- Governance-led impact and risk assessment methodologies
- Safeguards, controls, and accountability continuity mechanisms
- Data governance alignment including privacy, stewardship, quality, and lawful access
- Agentic AI readiness and delegated machine action controls
- Frontier-facing public authority and institutional hosting capacity analysis
- Leadership and oversight capacity strengthening
Institutional Scope
Ministries, regulators, oversight bodies, independent commissions, public-interest institutions, professional bodies, and selected regulated entities operating AI-enabled systems with public, legal, administrative, or systemic impact.
Core Governance Question
Can the institution lawfully host, supervise, interrupt, reconstruct, and remain accountable for the AI system once it enters public or regulated functions?
Digital Governance and Public Sector Innovation
Strengthening institutional readiness for accountable digital transformation.
Digital transformation requires clarity of mandates, coordination mechanisms, and supervisory authority. This programme strengthens governance foundations that enable coherent implementation and durable institutional trust.
AGCIH supports institutions to understand how digital and AI-enabled systems affect decision authority, accountability chains, institutional continuity, public trust, and the ability of officials to explain, review, and correct outcomes.
- Governance and readiness diagnostics
- Institutional design support including decision authority and escalation structures
- Leadership and oversight strengthening
- Governance, risk, and compliance integration
- Public-sector AI and digital transformation readiness
- Ecosystem-level coordination support
Governance Outcome
Institutional legitimacy, operational continuity, and accountability stability in digitally transformed environments.
Public Authority Focus
Digital systems should strengthen, not obscure, public authority. AGCIH helps institutions preserve decision rights, review pathways, escalation controls, and accountability chains as systems become embedded.
Research, Foresight and Civic Innovation
Evidence, anticipatory governance, and civic models that safeguard public value.
This programme strengthens institutional decision-making through applied research, anticipatory foresight, and civic governance models that identify emerging risks and reinforce inclusion, rights protection, and accountability.
AGCIH uses foresight to help institutions prepare for governance questions before systems become deeply embedded. This includes analysing how current AI adoption may evolve into agentic deployment, frontier-facing dependency, or institutional risk that exceeds existing oversight capacity.
- Applied governance research and institutional briefs
- Scenario planning and anticipatory risk studies
- Civic innovation models for inclusion and accountability
- Frontier-facing AI governance analysis for African public institutions
- Public trust and legitimacy assessments
Strategic Importance
Sustainable digital futures require governance systems that anticipate consequences early and maintain institutional legitimacy over time.
Foresight Question
What governance architecture must be built before AI capability exceeds ordinary institutional supervision?