7 Automation myths holding South African businesses back

In our context of South African workflow automation we hear some version of this in almost every first conversation: “We know we need to automate, but we’re not ready yet.” According to SASFA’s 2025 research on AI adoption among SA SMEs, more than half of South African small businesses plan to adopt AI-driven automation within the next twelve months, and most of them will not follow through. 

The reasons are not really about budget or ambition, but almost always about the stories that have built up around what automation is, how it works, and what it requires. After running operational audits across dozens of South African businesses, I can tell you that most of those stories are wrong.

These are the myths we encounter most often, and what is actually true in each case.

Myth 1: You need a developer to set up automation

A client we spoke to last year had been sitting on a list of processes she wanted to automate for almost eight months, waiting until she had the budget to bring in a developer. When we looked at what she actually needed, every single workflow on her list could be built using Make.com or Zapier, both of which have visual builders designed for business operators, not engineers. She had the budget. She had been spending it on the wrong assumption.

Workflow automation for South African SMEs has changed fundamentally over the past few years. The platforms that now handle most of what a small business needs: routing form submissions, creating tasks, sending notifications, logging records, triggering follow-ups, are built for people who understand their business processes, not people who can write code. Logic is configured visually, connections between tools are built through drop-down menus, and most common business tools already have pre-built connectors waiting to be used.

What does require real skill is the diagnosis that comes before the build: understanding which processes to automate, in what order, and how to structure them so they hold under the real conditions of your business. That is an operational problem more than a technical one, and it is where experienced input genuinely matters. The assumption that a developer is the starting point has stopped more businesses from automating than any actual technical barrier has.

Myth 2: South African workflow automation is too expensive

I find it useful to separate this into two distinct questions, because they get conflated in a way that makes the decision feel larger than it is.

The tools themselves are not expensive. Most of the platforms commonly used for SME workflow automation in South Africa run on monthly subscriptions (starting well below R500 at the time of writing this piece), scale with usage, and carry no upfront capital requirement. For context, that monthly cost is typically less than one hour of a senior employee’s time. 

The real investment is in the scoping and build work: understanding which processes to address, in what order, and constructing them correctly so they do not need to be rebuilt six months later. A focused, well-scoped automation covering one or two high-friction processes can be delivered in days. A more complex operational overhaul involving multiple systems and teams takes longer, and costs more.

The comparison that matters, though, is not the cost of automation against zero. It is the cost of automation against what your current manual process is already costing you in staff hours, compounding errors, and the founder’s direct involvement in tasks that should not require it. In most of the engagements we run, we find that the return on a targeted automation is measurable within the first quarter of operation. The expensive option, in almost every case, is the one where nothing changes.

Myth 3: Your data must be perfect before you start

This belief is one of the most effective ways to postpone automation indefinitely, and it is based on a misunderstanding of what most automations actually require.

We regularly assess businesses that have been sitting on automation plans for a year or more because their CRM has gaps, their spreadsheets are inconsistent, or their historical records are incomplete. When we look at what they actually want to automate, for example routing new enquiries, triggering invoice reminders or generating onboarding documents, the process needs four or five current data fields to function, not three years of clean history.

The question we ask is whether your data is adequate for the specific process you want to automate and the answer is almost always yes. There is also a practical case for starting before the data cleanup is complete: a process that runs through a system rather than through a chain of emails and a shared spreadsheet tends to improve data quality over time, because data capture becomes consistent by design rather than by habit. 

Clean data is often what good systems produce, not what they require.

Myth 4: Automating processes means losing control

This is understandable as a concern. Handing over something that currently runs through you or a trusted team member to a system, feels like stepping back from something important. In fact the businesses carrying the most operational risk are not the ones that have automated, they are the ones that have not.

A manual process is largely invisible. The logic lives in email threads, in informal agreements, in one person’s working memory. When something goes wrong, tracing it is slow and often inconclusive. When a key person leaves, the process can go with them, and in South Africa, where the ICT skills shortage grew from 14% in 2024 to 22% in 2025 and 89% of employers report that unfilled critical roles are directly damaging their operations, that is not a theoretical risk. It is a regular operational event for businesses of every size.

An automated process is visible by its own structure. Every step is defined, every trigger is recorded, every outcome is logged. You can see exactly what happened, when it happened, and at what point it broke. That level of visibility is more control, not less, and it is something most manual processes cannot provide. The question of which decisions stay with humans and which do not, is answered deliberately during the build stage, not left to chance. Automation should handle tasks that do not require human judgement. It should never replace humans on decisions that do, and any well designed implementation makes that boundary explicit.

Myth 5: Once automated, a process cannot be changed

This one comes from how software projects used to feel: months to build, expensive to modify, and risky to touch after launch. Modern workflow automation platforms were designed with a fundamentally different assumption: That business processes change, because they do.

Reconfiguring a workflow, whether that means adding a condition, changing a routing step, adjusting a trigger, or swapping out a connected tool, typically takes hours, not weeks. Because these workflows are built visually rather than in code, they are readable by anyone who understands the business logic, not only by whoever originally built them. That makes documentation straightforward and handovers manageable, which is relevant in an environment where the person who built something may not always be the person who maintains it.

A workflow we build for a client today will not look identical in two years, and it should not. Regulatory requirements shift, teams change structure, tools get replaced, the business evolves. Platforms are designed to absorb that. If you have encountered automation in the past that felt locked and immovable, that is most likely a reflection of how it was built or how poorly it was handed over, not a fundamental limitation of what automation can do.

Myth 6: You need to automate everything to see any benefit

Automation is sometimes presented as a transformation, which implies a certain scale of change before value appears. That framing causes businesses to set a threshold of readiness that keeps moving, and the automation never starts.

The businesses we have seen get the clearest, fastest returns from automation are not the ones that overhauled everything at once. They are the ones that identified one or two processes generating the most friction and automated those first. These usually involve repetitive data movement, manual follow-up, or a step that depends entirely on one person remembering to do something. 

A single automation that removes three hours of manual data entry per week delivers measurable value immediately and compounds over time. It also builds the team’s practical confidence in how automation works, which makes scoping the next intervention considerably easier.

The most sustainable approach is a sequence of targeted decisions, each one delivering a clear return, rather than a comprehensive transformation attempted in a single project. The goal at the start is straightforward: a working proof that the investment pays off.

Myth 7: Automation is only for manufacturing or tech companies

This misconception has its roots in the industrial meaning of the word “automation,” which conjures factory floors and robotics. That is a completely different category of technology from what a growing service business actually needs, and mixing up the two has cost South African SMEs real opportunities.

The kind of workflow automation that connects the tools a professional services business already uses, moves information between systems, removes manual copy/pasting, and makes sure client commitments are followed through consistently. It is built for exactly the businesses that most often dismiss it:  Accounting firms, legal practices, consulting businesses, estate agencies, HR consultancies and more.

 These are not edge cases for this technology; they are its primary market. The operational patterns that benefit most from automation are consistent across all of them: repetitive data movement between systems, multi-step approvals, follow-up sequences that depend on someone remembering, document generation, notification routing, billing triggers. If your business involves information moving between people and systems, which every service business does, workflow automation has direct applications for you.

The question is which processes, in your specific business, would produce the highest return if they ran without manual intervention. That is a business question, not a technical one, and it is exactly what a structured operational audit is designed to answer.

What actually needs to happen first

Automation is a business decision before it is a technical one, and the businesses that get the most from it are the ones that treat it that way. They diagnose before they build, identify which processes are genuinely costing time and money, determine the order in which to address them, and select tools that fit their team’s capacity and their operating environment rather than whatever tool was recommended in a webinar.

In South Africa, that operating environment matters in ways that generic frameworks do not account for. Lean teams, and compliance pressure are not exceptional conditions for local businesses, they are the baseline and automation that is not designed around them from the start tends not to last. We see this regularly, for example a solution built for a stable-infrastructure context, dropped into a South African business without adjustment, creating more problems than it solves within the first few months.

The Illuminate audit exists specifically to address this. It surfaces which processes, in your specific business, would produce the highest return if automated, and it does so before any software is recommended. With Lumino, the diagnosis always comes first.

Frequently Asked Questions