LeadAfrik flagship course
Become an AI-powered economist.
Not a generic “ChatGPT tips” course. This is how economists, analysts, researchers and policy professionals actually use AI to think better, analyse faster, and produce higher-quality work — from what is a token to shipping a real, verified deliverable.
The one habit that separates a professional from a prompt-monkey: verify everything. This whole course teaches it.
21 modules · capstone + verifiable certificate · first two modules free to preview
What you’ll be able to do
Finish able to produce, not just chat
Redesign your weekly workflow so AI does the drafts and you keep the judgement
Run a literature review in hours, not weeks — without a single hallucinated citation
Clean a messy dataset and pull defensible insight from it with AI
Get AI to explain, debug and interpret an econometric model — and catch its confident mistakes
Write a decision-ready policy brief and a real investment memo, every figure traced to source
Build your own AI economist toolkit: custom assistants, a prompt library, a grounded knowledge base
What’s inside
A complete programme, not a playlist
21 modules, 5 parts
Foundations → applied craft → capstone
63 quizzes + 21 exercises
Practice in every lesson, instant feedback
AI tutor in every module
Mwalimu explains, checks your work, answers back
A reusable prompt library
Battle-tested economist prompts you can copy today
Capstone + verifiable certificate
Ship a real deliverable; earn a certificate employers can check
African by default
KNBS, CBK, NSE, county budgets — real local data
The curriculum
21 modules, from foundations to a shipped deliverable
Keep the rigour of how AI actually works; add the applied workflows that turn it into real output.
Part 1
Foundations — understand AI
- ◆The AI economist's mindset
- ◆What AI is (and isn't) for an analyst
- ◆The machine-learning toolbox
- ◆From OLS to neural networks
- ◆NLP foundations — tokens, embeddings, attention
- ◆How large language models actually work
Part 2
Working with AI — the craft
- ◆Prompt engineering for analysis
- ◆Retrieval-augmented generation (RAG)
- ◆AI agents and tool use
- ◆AI for analyst productivity
Part 3
The economist's craft — produce real work
- ◆AI for economic research
- ◆AI for data analysis
- ◆AI-assisted econometrics
- ◆Coding with AI: the analyst's assistant
- ◆AI for policy analysis
- ◆AI for finance & business
Part 4
Judgement & the big picture
- ◆AI in finance and economics — real use cases
- ◆Evaluation, hallucination, and safety
- ◆The economics, ethics, and geopolitics of AI
Part 5
Build your toolkit → capstone
- ◆Building your AI economist toolkit
- ◆Capstone — ship a real deliverable (+ certificate)
The capstone
You finish with something real
The course ends with a graded capstone. You choose one: a full economic research report, a decision-ready policy brief, or an interactive economic dashboard — built with AI, and defensible line by line.
The rubric weights rigour and verification at 55%. A single hallucinated statistic fails the piece — exactly as it would end your credibility in a real job.
A certificate you can prove
Pass the AI for Economists & Analysts assessment and your capstone, and you earn a LeadAfrik certificate with a unique ID — downloadable as a PDF and verifiable by any employer online. A claim you can defend, because you did the work and checked it yourself.
See the certification assessment →Included: the prompt library
Battle-tested economist prompts — copy and use
A taste of the reusable prompt library you build in the course. Each one bakes in the verification discipline. Copy, drop in your context, and go.
Extract a paper's methodology
You are a careful research assistant. From the paper text below, extract ONLY what is stated — do not add outside knowledge. Return: (1) the research question; (2) the data and sample; (3) the identification strategy / method; (4) the main findings with the numbers; (5) the stated limitations. For anything not in the text, write "NOT STATED". Quote the sentence that supports each point. PAPER TEXT: [paste]
Stress-test an identification strategy
Act as a hostile but fair referee for an economics seminar. Here is my identification strategy: [describe your design, data, and the causal claim]. List, in order of severity, the threats to identification (omitted variables, selection, simultaneity, measurement error, bad controls). For each, say how damaging it is and what evidence would rule it out. End with the single change that would most strengthen the design. Do not reassure me — find the holes.
Profile and clean a dataset
You are a data analyst. First PROFILE this dataset, do not change it yet: list every column, its type, the count of missing values, obvious anomalies (sentinels like -99, mixed types, "7.1%" as text, inconsistent category spellings), and any duplicate rows. Then propose a cleaning plan as a numbered list. For each step, say exactly what changes and why. Wait for my approval before writing code, and keep a change-log of every row/value you alter. DATA: [paste CSV or describe the file]
Draft a policy-brief section (grounded)
You are drafting the EVIDENCE section of a policy brief for [decision-maker]. Use ONLY the sources pasted below — do not introduce any statistic, date, or claim that is not in them. For every figure you use, cite the source in brackets and quote the exact phrase it came from. If a needed number is missing, write "[NOT IN SOURCES — verify]" rather than inventing one. Write in plain, decision-ready prose, ~300 words. SOURCES: [paste the budget statement / report excerpts]
Structure an investment memo
Help me structure an investment-research memo on [company, ticker, exchange]. Produce the skeleton with these sections: thesis (one paragraph), business & drivers, financials (key ratios to compute), valuation approach, the three biggest risks, and a recommendation. For each financial figure, add a placeholder [SOURCE: filing, page] — I will fill and verify every number against the primary filing myself. Do NOT state any specific financial figure as fact; flag where I must look it up.
Fact-check my own draft
You are a fact-checker. Below is a draft I wrote (possibly with AI help). Go through it and flag: (1) every specific statistic, date, or named figure that needs a source; (2) every citation that must be confirmed to exist; (3) any causal claim made from correlational evidence; (4) any sentence that sounds confident but is unsupported. Return a numbered checklist of what I must verify before this ships. Do not fix it — just find what could be wrong. DRAFT: [paste]
Who it’s for
Economists and analysts who don't want to be left behind by AI
Researchers and postgraduate students who write and analyse for a living
Policy and public-sector professionals producing briefs and evidence
Finance and consulting analysts who ship memos and models
Anyone who has tried ChatGPT and thought “there must be more to this”
Employers upskilling a whole team — ask about cohort access
Start today
A programme worth far more than it costs
The full flagship — 21 modules, the prompt library, the capstone and the certificate.
First two modules are free to preview. Standalone KES 4,999, or get this and every other LeadAfrik course with all-access at KES 6,000— often the better deal.