Home / Services / Business Systems / The Operating System

The Operating System

Stop running your business on a spreadsheet nobody fully understands.

R165,000 – R300,000 Fixed price · delivered in 6–10 weeks · optional care plan R7,500–R25,000/mo from R16,500/moover 12 months · R198,000 totalDelivery is still 6–10 weeks — this only spreads the payment, it doesn't stretch the project.

Show price as

Is this you?

Your business runs on a spreadsheet nobody fully understands

There's a file. It might be a Google Sheet with eleven tabs, or a no-code database somebody set up enthusiastically and then left. It holds something important, jobs, stock, clients, money, and it has grown one column at a time for years.

Three things are now true about it. Nobody fully understands all of it. Two people maintain slightly different versions. And every month somebody spends a day and a half turning it into a report, by hand, that is out of date by the time it's sent.

You've probably already tried to fix this. Maybe you bought a SaaS tool that did 70% of what you needed and none of the specific thing that matters. Maybe you added an AI assistant that gives confident answers about data it can't actually see. The reason those didn't work isn't that they were bad tools, it's that the underlying data was never modelled properly, and no tool fixes that from the outside.

What you
actually get

Every line below is a deliverable, not an intention. If it's on this list it's in the proposal, and if it's not on this list it isn't included.

Data architecture designed around your actual business

Not tabs. Proper tables with relationships, rollups and validation, modelled on how the work really flows, including the exceptions everyone currently handles by remembering them.

Migration with your history intact

Years of records moved across and reconciled, not abandoned. If the old numbers don't add up after migration, that's a bug and I fix it. You keep the past.

Three to five automated workflows

The specific recurring jobs eating your week, a report that assembles itself, an enquiry that routes itself, a reconciliation that runs overnight. Chosen with you, by time saved.

AI capture: voice or chat in, structured records out

Say what happened, or paste an email, and it lands as a properly structured record in the right table with the right fields. This is the part that makes people actually use the system instead of updating it on Friday from memory.

A live dashboard

One screen showing the state of the business, derived from the data rather than retyped into a deck. It's correct because it can't be anything else.

Role-based views

Each person sees what they need and nothing they don't. Fewer mistakes, less noise, and one less reason for someone to keep their own private copy.

SOPs and training, written for humans

Documentation your team can follow without me, because a system only one person understands is the problem you started with, wearing a nicer interface.

How it runs

So you know what's happening and when, and so nobody has to ask for a status update.

Weeks 1–2

Discovery and modelling

I sit with how the work actually happens, not how the process document says it does. Output is a data model you sign off.

Weeks 3–5

Build and migrate

Tables, relationships, validation, then your history moved across and reconciled against the old source.

Weeks 6–7

Automation and AI capture

The workflows, the capture layer, the alerting. This is where the hours start coming back.

Week 8

Dashboards and views

The reporting layer, plus role-based views per person.

Weeks 9–10

Training and handover

Your team trained on it, SOPs written, and two weeks of me watching for the things that only show up in real use.

+60 days

Support

Fixes and tuning included. Real usage always surfaces something the discovery didn't.

What I need from you

The projects that run late almost always run late for one of these reasons, so they're worth reading properly before we start.

Access to the systems and files the data currently lives in
One person who genuinely knows how the work happens, usually not the most senior person in the room
Two to three hours a week during discovery, and less after that
Honesty about the workarounds. The undocumented ones are the ones that break the build if I don't know about them
A decision on who owns the system internally once it's live

What's not included

On the public page rather than buried in the proposal, so there's no argument in week five.

Custom-coded application (that's Enterprise)
Unlimited workflows
Platform licence costs
Ongoing changes after 60 days

Your other options

What happens if you don't do this

Keep the spreadsheet. It works, right up until the person who understands it leaves, or two versions diverge far enough that a decision gets made on the wrong number. The risk isn't that it stops working, it's that you don't find out it stopped being accurate.

Buy a SaaS tool. Sometimes correct, and I'll say so if it is. But you'll be shaping your business around someone else's model of it, paying per seat forever, and you'll still need the glue between it and everything else. The glue is usually the actual problem.

Start smaller. Entirely reasonable. Take one process, automate it for R35,000–R65,000, and see whether the approach pays for itself before committing to the whole thing, Automate One Process exists precisely for this. Most people who buy the Operating System bought that first.

Straight
answers

Why the R135,000 spread between R165,000 and R300,000?

Number of workflows, complexity of the data model, how messy the migration is, and how many people need training. After the first discovery conversation I can usually place you within R30,000, and the fixed number is agreed before anything starts.

Is this just a no-code database? I could set that up myself.

You could, and some people should. What you're paying for is the modelling, the part that decides whether the thing still works in two years or collapses into tabs again. Most self-built bases fail on relationships and validation, not on features.

What does it cost to run afterwards?

Platform licences, which you pay directly and which are usually a few hundred rand a month per seat. I don't mark those up or resell them. The optional care plan is R7,500–R25,000 a month if you want it maintained and extended.

What if we outgrow it?

Then you've got a properly modelled dataset, which is exactly what you'd need to move to a custom platform. That's the Enterprise tier, and the work here isn't wasted, the data model is the expensive part and it carries across.

You're one person. What if you disappear halfway through?

The build is phased and you own everything at each phase, not at the end. Documentation is written as we go rather than at handover. If I vanished after week five you'd have a working data model and migrated history that another developer could pick up, which is not true of most agency projects at week five.

Need a real interface with permissions for a team?

Custom Internal Platform, Discovery R45,000 – R75,000

When the spreadsheet became a business system, this is the replacement.

See Custom Internal Platform →

Sound like
your problem?

One call, no deck. If this isn't the right package I'll tell you which is.

Email instead