institutional reform

Government capacity constraints are rarely about individual competence and frequently about systems that lose knowledge between administrations, duplicate work across departments, fragment digital services across hundreds of disconnected portals, and leave staff without the tools or documentation to do their jobs. These briefs diagnose specific institutional dysfunctions and propose interventions that build cumulative capability rather than one-off fixes.

public policy lab · policylabmv.com · @policylabmv

What's here

Organizational efficiency diagnosis Knowledge management repository Institutional memory preservation Pro bono capacity support model Standard prompter AI templates Digitalization unified architecture AI for government LLMs + regulation

How the pieces connect

A connected system for institutional capability

01

Diagnose

Information loss, duplication, and fragmentation as root causes of dysfunction

02

Document

Institutional memory preservation and knowledge management repositories

03

Digitalize

Unified digital architecture replacing fragmented portals

04

Augment

AI templates and standard prompters encoding best practices

05

Support

Pro bono capacity filling gaps the system cannot fill itself

Policy briefs

Diagnosis · systems analysis · ◷ 9 min read

Organizational efficiency

How information infrastructure gaps cause policy failure across government. Knowledge that exists in one department is invisible to another; work done under one administration is repeated under the next; staff spend hours producing documents that could be templated. The diagnosis that frames the interventions in every other brief on this page.

Full analysis →

This brief connects to everything else here: the knowledge repository addresses the storage problem, institutional memory addresses the transition problem, digitalization addresses the access problem, the prompter addresses the production problem, and pro bono support addresses the capacity problem. Each intervention makes the others more effective.

Knowledge infrastructure · ◷ 4 min read

Knowledge management repository

A centralised searchable digital repository of policy documents, research, news archives, and literature reviews. Prevents the cycle where every new team spends months doing background research that previous teams already completed, because that work was stored in email inboxes and personal drives.

Continuity · documentation · ◷ 4 min read

Institutional memory preservation

Comprehensive documentation of programmes and policies across all departments so that institutional knowledge survives staff turnover and political transitions. What was tried, what worked, what didn't, and why – recorded in a form that the next team can actually find and use.

Capacity · implementation · ◷ 4 min read

Pro bono capacity support

Ongoing pro bono technical support relationships that fill specific capacity gaps in public sector agencies. Case study: how sustained technical support helped the National Library secure MVR 270 million in investments it could not have pursued without external expertise.

AI tools · templates

Standard prompter

Form-based AI prompting templates that encode institutional best practices for common government documents – cabinet papers, UN reporting formats, policy briefs, analytical reports. Quality becomes a function of the template design rather than individual skill, making strong outputs accessible to any staff member.

Digital infrastructure · ◷ 13 min read

Digitalization with purpose

The Maldives has built hundreds of disconnected government portals – citizens still carry physical documents between offices because the systems don't talk to each other. The case for a unified national digital architecture built by an elite central unit, replacing fragmented digitisation with coherent integration.

AI & technology in government

Understanding and governing AI

Intro to LLMs for policymakers · ◷ 55 min read

How large language models actually work – next-token prediction, pattern matching, the difference between statistical fluency and genuine reasoning. Written for policymakers who need to make informed decisions about AI deployment without a computer science background.

Human rules in an AI world · ◷ 24 min read

The regulatory architecture needed for AI-driven harms: deepfakes used for political manipulation, algorithmic bias in automated decisions, fabricated evidence in courts, voice cloning for fraud. Human-in-the-loop requirements and maintaining accountability chains when systems make consequential decisions.