What is EDIS?
An integrated economic development intelligence platform that connects data, models, GIS and AI agents to help governments and institutions analyse policies, allocate resources and make evidence-based development decisions.
A digital economic planning and investment adviser
Governments continuously decide where to build roads, hospitals and schools; where to expand electricity, water and internet connectivity; which agricultural and industrial investments to support; and how to respond to climate and other development risks.
The information needed to make those decisions usually exists. It sits across the finance ministry, the planning authority, the statistics bureau, sector ministries and local governments. The difficulty is that data, budgets, economic models, sector plans, maps and individual projects are analysed through separate systems, on different assumptions, at different times.
EDIS brings them into one place and makes them answer a question together. A policymaker asks where the next hospitals should go, or how an additional trillion dollars should be allocated. EDIS routes that question across the relevant data, models, maps and appraisal tools, runs them in sequence, and returns options, locations, costs, expected benefits, risks and the evidence behind each one.
The final decision remains a human and political responsibility. EDIS makes the evidence for that decision explicit, comparable and reproducible.
In one line
EDIS — the Economic Development Intelligence System — brings the models, data and analytical infrastructure together, and is the platform through which governments, institutions and analysts use them.
What EDIS combines
- Economic and fiscal models
- Sector planning models across sixteen sectors
- Geographic information and spatial analysis
- Public finance, budgets and unit costs
- Optimization and investment appraisal
- Scenario analysis and forecasting
- AI agents for routing, explanation and reporting
- Monitoring, evaluation and learning
Questions EDIS is built to answer
Each of these is a registered question class with a defined agent, workflow, model set, data requirement and validation status.
Public finance
- What happens if government increases health spending?
- What happens to GDP if VAT is reduced?
- How should an additional $1 billion budget be allocated?
- What are the fiscal implications of expanding universal health coverage?
- What is the optimal public investment portfolio?
Infrastructure and services
- Which roads should be prioritised?
- Where should new hospitals be located?
- Which districts require additional schools?
- What energy investments are required by 2040?
- How should energy, transport and industrial investment be coordinated?
Growth, trade and climate
- Which industries should a country prioritise?
- How will an AfCFTA tariff scenario affect trade and employment?
- How will climate change affect agriculture?
- How can government reduce poverty most efficiently?
- Which regions are falling behind, and on what?
Model-driven intelligence, not a language model with opinions
EDIS is deliberately not positioned as a chatbot. The AI layer understands, routes, orchestrates, explains and reports. The EDIS analytical layer calculates, simulates, optimizes, forecasts, maps and validates. The separation is architectural.
What the AI layer does
- Interprets the question and detects country, sector, scenario and horizon
- Selects the registered agents that accept this class of question
- Constructs the model workflow and its execution order
- Monitors the run and reports failures honestly
- Explains the results and drafts the report around them
What the AI layer never does
- Estimate a GDP, revenue, poverty or cost figure
- Fill a gap where a dataset is missing
- Override a model that has not converged
- Publish a number without model version and data source
- Grant itself access to data the user is not licensed for
Six rules the platform is built on
Data first
No model executes against an unverified dataset. Every dataset carries provenance, metadata, a quality score, a version, an update history and licensing information.
Modular design
Every subsystem operates independently and communicates only through defined interfaces and shared schemas. No subsystem reaches into another’s internals.
Model interoperability
Macro-fiscal, CGE, sector, infrastructure, climate, microsimulation and optimization models exchange standardised outputs along a common chain.
AI augments analysts
AI summarises, routes, explains, retrieves, drafts and checks. Policy judgement remains a human responsibility, and the platform is built to make that division visible.
Cloud native
Containerised services, background workers, queued model runs, horizontal scaling, continuous deployment and fault tolerance, so long simulations never block the interface.
Enterprise quality
Documentation, testing, logging, security, performance and maintainability standards are met before any component is considered complete.
Built for a continent, demonstrated in one country
Country, region, district, year and scenario are first-class dimensions throughout the platform. No country is hard-coded. Utopia is the initial demonstration country; the architecture is designed to extend across African economies, with multi-country comparison for regional institutions and development partners.
Country
Model instances, data, licences and permissions are all scoped by country.
Region and district
Sub-national results throughout, with spatial analysis at settlement level where data allows.
Year
Baseline year and forecast horizon are set per run and recorded in the provenance.
Scenario
Every result belongs to a named, saved, comparable scenario with a run ID.