A reality check for the person running the numbers at a South African SME.
Let’s be honest about how this actually works.
You run a R5m or R20m business, you’re smart, you’ve got someone handling the processing — a bookkeeper, maybe a small internal team. You’ve got an accountant you’ve been with for years who does the tax. The cost of this team gets a decent share of your monthly budget. And when something feels off, or you need a second opinion on a big call, you’ve started doing something you probably wouldn’t have admitted to two years ago.
You ask ChatGPT, or Claude. Or Gemini. Or whichever one your browser opens first.
You type in your question — “what’s a reasonable gross margin for a services business at my revenue level”, or “how should I think about this intercompany loan”, or “is this VAT treatment correct” — and you get an answer in thirty seconds that would have taken your accountant three days and possibly a sizable invoice.
This isn’t a criticism. It’s just what’s happening. Everyone is doing it.
What AI got right about finance — and what changed in early 2026
AI has been creeping into finance for years. But something significantly shifted in early 2026.
The models got better at reasoning through numbers. Agentic AI — systems that don’t just answer questions but actually take action, run processes, and check their own work — moved from experiment to deployment. Global CFOs, surveyed at the start of the year, stopped talking about AI as a future investment and started describing it as a current team member. The AI accounting market is on track to hit a multiple of billions of rands in 2026, with SME adoption growing faster than any other segment.
In short: the technology is here, it works, and the businesses using it are pulling ahead.
Here’s the part nobody is talking about though.
Your accountant — whether that’s an internal hire or an external firm — is already using AI. They’re using it to process faster, reconcile quicker, draft reports in a fraction of the time. AI has made their operation significantly more efficient.
And yet your monthly fee hasn’t moved.
The productivity gain went to them. Not to you.
The three ways SMEs actually run their finances
When I talk to business owners and finance heads at companies, the setup is almost always one of three things.
The DIY model. The owner manages the books directly — often in a spreadsheet, sometimes in Xero, Sage or a free online accounting tool, with AI tools filling the gaps when something tricky comes up. Tax gets handled separately by an accountant who sees the numbers once a year. It works well enough. Until it doesn’t.
The internal team model. There’s a bookkeeper, maybe a financial manager. They process transactions, produce month-end reports, keep things ticking. The owner gets a report, glances at it, and moves on. The accountant handles compliance. Everyone assumes the system is working.
The outsourced model. Processing and reporting goes to an external firm. Variable quality, variable responsiveness. The owner often feels disconnected from their own numbers.
All three have the same problem, just expressed differently. Processing is happening. Insight isn’t.
Processing is not the same as insight and insight is not the same as intelligence
What I’ve personally seen with over a decade of CFO experience, managing the finances of companies of all sizes and jurisdictions: the companies that compound don’t have more data. They have better interpretation of the data they already have.
Your finance team, internal or external, can process. They can reconcile the bank, post the journals, produce the trial balance. The question worth asking honestly, besides having a higher cost than AI, is whether anyone is connecting the dots beyond the numbers themselves.
Because your data is telling a story right now. Your margin, your revenue concentration risk, your working capital cycle. The three clients who are quietly making you money and the two who are quietly destroying it.
That story exists in your numbers today. Is anyone reading it?
The real case against DIY
I’m not going to tell you that doing it yourself is wrong. If you’re a financially literate owner running a lean, non-VAT registered operation, doing your own books is a completely rational decision. Keep your costs low. Use a good AI tool when you need a second opinion. Get your accountant in at year-end.
But I want to push back on two assumptions: that your time is free, and that the opportunity cost of DIY is zero.
Firstly, although it may feel like your time is free, it isn’t. Your time has a cost — and it’s probably the highest cost in the business. Every hour you spend reconciling transactions or trying to figure out why the VAT doesn’t balance is an hour you’re not spending on the thing that actually grows the business. The question isn’t whether you can do it. It’s whether that’s really where your time should go.
Secondly, every piece of financial information you DIY is a valuable piece of data for your business. If you’re not making sure it’s processed in a standardised manner and further analysed for value beyond just your revenue number and cash balance, there’s a real opportunity cost.
The thing AI can’t do for you
Use AI tools. They’re genuinely useful, and we’d be dishonest if we told you otherwise.
AI can answer the question you know to ask. What about all the questions you haven’t thought to ask or prompted? It can’t tell you the question you should have been asking three months ago.
That’s what we do. And it’s why the data insight piece isn’t an add-on at Vedant — for Complete clients, it’s the whole point.
What Vedant actually offers — and why it’s priced the way it is
We built Vedant on a simple premise: AI should make professional finance cheaper for you, not just more profitable for the accounting industry.
We use AI across our entire operation — processing, reconciliation, reporting, anomaly detection. It makes us faster and more accurate. And we pass that efficiency directly to clients through our pricing.
The Essentials package is for the business that needs clean, professional books without the overhead. Cloud-based processing with no frills or insights to ensure the cost is kept as low as possible. Priced for businesses that value cost over complexity — particularly owner-managed, non-VAT registered operations where accurate processing is the primary need.
The Complete package is for the business that wants to actually use its numbers. Everything in Essentials, plus management reporting with real commentary and analysis, cash flow visibility, the financial narrative that tells you what’s happening in your business and why. This is where the insight lives. If you’re VAT registered, growing, and making decisions that depend on understanding your numbers month to month, Complete is the right level.
Beyond Complete — bespoke CFO-level advisory, financial modelling, deal support — exists for businesses that need more than a finance function. That’s scoped and quoted separately, case by case.
One way to test it
If you want to see what your numbers are actually saying — right now, with no commitment — upload your trial balance to the Vedant Report. You’ll get an AI-powered financial analysis in minutes: margins, ratios, red flags, and commentary.
It’s free. It takes three minutes. And it will tell you, pretty quickly, whether your current setup is giving you the story your business is trying to tell.