A practical, self-paced challenge that teaches accountants and bookkeepers to put AI to work on real jobs, and to catch it when it is confidently wrong.

Drafting, summarising, categorising, explaining the numbers: AI does all of it in seconds, and some of it wrong in ways only a trained professional will catch. The fluency is the trap. The tool will write a beautiful paragraph explaining a profit the client never made, and never flag a doubt.

Over 28 days, one short task at a time, you learn to direct AI like a junior and to verify what it gives back. Every day pairs an AI task with the judgment check that matters more, because for an accountant the judgment is the job.

A letter from the CEO

To the accountant or bookkeeper reading this,

There is a quiet anxiety running through our profession right now, and it goes something like this: will AI replace me? I want to answer it directly, because the honest answer is more useful than the comfortable one.

No. AI will not replace you. But it will replace the accountant who refuses to learn it, with one who has. And it will do something else that worries me more: it will flood the market with confident, fluent, unverified financial output. Numbers that look right. Reports that read beautifully. Some of them wrong in ways only a trained professional will catch.

That is the opportunity. The world is about to need people who can tell when the machine is wrong far more than it needs people who can operate the machine. That person is you. That is not a consolation prize, it is the most defensible position in the entire economy.

The large firms understand this. Deloitte has put Claude in front of 470,000 employees. EY has built 150 AI agents for its tax staff. They are spending billions, and they are quietly cutting their graduate intake because they expect AI to do the junior work. The accountants serving South Africa’s small businesses do not have a billion-rand budget. What they have is judgment, client relationships, and the trust of the businesses that keep our economy alive.

CIBA’s whole reason for existing is to make sure those accountants are not left behind. When your clients grow, your income grows, and the economy grows with them. An accountant who uses AI well keeps more SMEs compliant and alive. That is not abstract. That is GDP, one business at a time.

So we built this guide and we are giving it away. Not as a sales funnel. As a contribution to a profession we believe in. Work through it honestly, keep your judgment sharp, protect your clients’ data, and you will come out the other side as the professional the market is short of.

You are not a spectator in this economy. You are one of the people shaping it. Let’s make sure the rest of the profession sees it too.

Nicolaas van Wyk
Chief Executive Officer
Chartered Institute for Business Accountants NPC (CIBA)

Who it is for

Accountants and bookkeepers in South Africa, in practice or employed, who serve small and mid-sized businesses. It assumes no AI experience. It assumes you already know the accounting; the challenge teaches you to apply AI to it safely.

What you will be able to do

• ✔  Use AI to draft, summarise, categorise and explain financial information, then verify every output before you trust it.
• ✔  Spot the three ways AI fails an accountant: confident wrong figures, invented detail, and citations to rules that do not exist.
• ✔  Have AI run exact, code-based calculations instead of guessing at the numbers.
• ✔  Turn AI-assisted work into services you can charge for: plain-English summaries, commentary, KPI dashboards and insights notes.
• ✔  Protect client data and meet your POPIA obligations whenever you use an AI tool.

What is inside

• ✔  Thirty-one short lessons: an orientation, the 28 daily tasks, and a data-security annexure.
• ✔  A clear arc across four weeks: protect yourself, serve the SME, grow the practice, prove your value.
• ✔  A judgment check on every day, the specific things to catch before you trust the output.
• ✔  A ready-to-use prompt for each task, written for South African accounting work.
• ✔  Built and tested with Claude, and written to work with any capable general AI.
• ✔  Self-paced and online, free and open to members and non-members alike.

Course contents

Week 1 · Protect yourself

Week 2 · Serve the SME

Week 3 · Grow the practice

Week 4 · Prove your value

Self-paced, about 15 to 20 minutes a day over four weeks. An orientation lesson and a data-security annexure sit around the 28 days.

How it works

• ✔  Self-paced and online, one short task a day for 28 days.
• ✔  Start with the orientation, then work the days in order; each one builds the habit.
• ✔  Read the judgment check every time. The day you stop verifying is the day the risk becomes real.
• ✔  Read the data-security annexure before you put any real client information into an AI tool.
• ✔  Free and open: no membership required.