Automate The Right Process First, Not The Loudest One
Most AI automation initiatives at Johannesburg SMEs fail not because the tool was wrong, but because the first process automated was whichever one a manager complained about loudest. It hits its own exceptions within weeks, and that stumble becomes the internal case against AI. Our AI consulting work scores every process first, before any tool gets picked.
Five Ways Johannesburg SMEs End Up Automating The Wrong Process First
The Same Three Patterns, Every Time
By the time a business calls us about AI, someone has usually already picked a process to automate. In scoping conversations, the same three patterns keep showing up, and all three explain why a first attempt stalled before it proved anything.
The Process Everyone Complains About
This is usually the invoicing run mentioned in every meeting, or the checklist ops has flagged for two years running. It gets picked first because it is visible, and everyone agrees it is painful. But it's the complicated ones that actually break the build.
The Process Nobody Has Mapped End To End
This one usually touches client records, data, or files that have never been checked for what an AI tool could do once it is connected. Before scoring a process for a roadmap, know what's exposed, which is exactly what an AI readiness assessment is for.
The Process Still Changing Every Month
Sequencing a process that is still being redesigned means a dated workflow before the build finishes. We ask every client one question: has this been stable for at least a quarter? If the answer is no, it waits, because automating a moving target wastes the build.
None Of These Rule Automation Out
Every one of these is fixable with the right sequencing, not a reason to shelve the idea. Book a process automation strategy call and we will tell you honestly which pattern your business is in.
Most businesses don't fail at automation because the tool was wrong. They fail because the process underneath it was still moving: a workflow half-redesigned, dependencies nobody mapped, an invoicing run everyone agreed was painful but nobody checked end to end. As your process automation strategy partner, we score the sequence before we touch the build, so the first automation you ship is the one that was actually ready, not just the one that got picked first.
How We Score Every Process Before It Makes The Roadmap
A roadmap that just lists whichever processes got mentioned in a meeting is not a strategy, it is a wish list. Before anything is sequenced, every candidate process goes through the same four checks.
Two Things You Get Out Of This Phase
Two documents come out the other end of this phase, and both hand straight to whoever builds or integrates the automation next, so nothing gets lost in translation between scoping and delivery.
Questions Every Candidate Processes
How often does it run? How rule-based is it? How clean and accessible is the data feeding it? What does it currently cost in hours? A process that scores well on all four moves forward, one that does not gets parked, not forced through.
Plotting Effort Against Impact
Every scored process gets placed on an effort versus impact view, so the contained, high volume, rule based win is visible next to the ambitious multi quarter build, instead of picking whichever one is loudest internally.
Discovery With The People Who Run It Daily
We sit with department heads and the staff who actually do the work, not just their managers, because the real exceptions and workarounds usually live at that level. A first discovery pass like this typically surfaces
A Roadmap Sequenced By Evidence, Not Opinion
Every process that makes the list carries its scoring with it: frequency, the rule versus judgment split, data readiness and hours saved. Whoever picks up the build can see exactly why process one comes before process four.
Where Tool Selection Comes In, Not Before
The roadmap tells you which process goes first and why. It doesn’t pick the platform. Once a process is sequenced, choosing the right tool is a separate, scored step, which is what AI tool selection and integration covers next.
What You Actually Get When The Roadmap Is Signed Off
Nothing Vague Gets Handed Over
A roadmap on its own is a list of names in order. What actually gets handed over, per process, is two working documents, not a slide deck summary. Whoever builds or integrates the automation next works from these.
A Reality Based Workflow Map
How the process runs, step by step, affects how it will run once automated, so nothing gets lost between the person who knows the process and the builder. This is the document that turns an idea into a buildable spec.
A Documented Business Case
Current hours spent, the realistic time or cost saved once it is automated, and a payback estimate, calculated before a budget is added.
Where This Leads Next
A sequenced, costed roadmap is the input the rest of our Artificial Intelligence services build from, whichever process goes first.
Why We Don't Publish A Flat Roadmap Fee Upfront
There’s no flat fee for a process automation roadmap: scoping a five person finance team looks nothing like scoping a forty person logistics operation running three disconnected systems. What actually moves the price is how many candidate processes are worth assessing, how tangled the exceptions inside each one are, how much of the current workflow already exists on paper instead of in someone’s head, and how many departments need a seat at the discovery workshops.
01 How Many Candidate Processes Need Assessing
02 How Tangled Each Process's Exceptions Are
03 How Much Of It Is Documented Already
04 How Many Departments Need To Attend
Who This Is For
This fits Johannesburg businesses with at least one process running weekly or more, a named person who can walk it through even without documentation, and data sitting somewhere, spreadsheets and email included. It's the wrong stage for a one-off task, or a workflow still changing shape month to month, since sequencing it now locks in something about to move.
How it Works
frequently asked questions
How long does the strategy phase take before we get a roadmap in hand?
Most engagements move from the first discovery workshop to a signed off roadmap in, covering process scoring, an effort versus impact view, and a business case for each shortlisted process.
box before the real cutover, specifically to avoid this. Some brief, planned downtime may still be needed depending on how much legacy data has to move.
Do you build the automation yourselves, or does this phase only produce the plan?
This phase produces the sequenced roadmap and the business case behind it, not the build. Once a process is chosen and its business case approved, most Johannesburg clients move straight into scoring and selecting the automation tool itself.
What happens if none of our processes are documented anywhere right now?
That is the normal starting point, not a blocker. Discovery workshops with your department heads and process owners are how the first documented as is workflow gets built, and most SMEs we have scoped had nothing written down before that session.
How do you decide which process goes first when three managers each think theirs should be first?
Every candidate process is scored the same way: how often it runs, how rule based versus judgment based it is, how clean the underlying data already is, and what it currently costs in hours. Plotting those scores on an effort versus impact view takes the office politics out of the decision.
You Have The AI Ambition, Now It Needs A Sequence
Every Johannesburg SME we scope has more automation ideas than budget, and the roadmap decides which one gets built.