The method behind MyAutomates

Ask the people who do the work. Build only what they justify.

Five steps, in the same order every time: listen to every person, measure what the inefficiencies cost, decide what each one needs, build one system at a time, and prove it worked.

// People first · Evidence second · Software last

The method starts with people, not software.

Nothing here replaces your team. The people doing the work are the source of every finding, the reason for every system we build, and the people who use what we build. The work that goes is the retyping, the chasing and the waiting. The people stay.

00Where it fits

The system is never the starting point.

Most AI projects start from a tool someone saw, or from a manager's impression of where the time goes, and the larger the business, the further that impression sits from the work. Both skip the people who actually do the work, which is where the detail is.

So we start there. The interviews find the work. The measurements put a cost on it. Only then do we decide what to build, and only what the evidence supports. After that, we measure again.

Listen to everyone. Measure the cost. Build only what the evidence supports.

01Why it exists

The expensive part of AI is building the wrong thing.

Ask leadership where the time goes and you get the business as it is written down. Ask the person doing the task and you get the business as it actually runs: the report rebuilt by hand every Friday, the approval that waits two days for one signature, the customer question answered from scratch for the hundredth time.

The gap between those two versions is where the time and money leak. It is also where most AI projects go wrong. A system specified from the written-down version automates the wrong step, or the right step in the wrong way, and the team quietly goes back to the old way of doing it. Nobody uses it, and the money is gone.

The software itself is rarely the problem. The tools available today are good and getting cheaper. What no tool can tell you is which of your problems is worth solving first, and what solving it would actually save. Only the people inside the business know that, and they will only tell you if they are asked properly.

There is a second reason. People who were asked what slows them down are far more likely to use the thing that fixes it. A system nobody uses has failed however well it was built, so the interviews are as much about adoption as discovery.

Right first, then fast.

02The five steps

Each one fixes a problem you have probably already had.

// 01

Listen

Every person

Part of The Audit

"We asked the team in a meeting. Nobody said anything useful."

Every person in the business gets a private interview, 30 to 45 minutes, one at a time, on site. We ask about the actual day: what they do, what they repeat, what they wait on, and the workarounds nobody wrote down. It is recorded with their written consent, and nothing they say is attributed to them. People describe the job as it really runs when they know their name will not reach their manager.

// 02

Measure

In hours

Part of The Audit

"Everyone agrees it is inefficient. Nobody can say what it costs."

Every inefficiency the interviews surface is put in hours per week, and those hours are priced at your own cost of labour. A complaint becomes a number you can rank. The same numbers become the baseline that everything we build is later measured against.

// 03

Decide

Three answers

Part of The Audit

"We bought the software and nothing changed."

Each finding gets one of three answers: automate it, change the process, or stop doing it. Software is one answer in three, and often not the right one. The findings are then put in order of what they cost against what they take to fix, and that order is the roadmap. It is written so that another developer could work from it, which keeps us honest about what we recommend.

// 04

Build

One at a time

Part of The Build and Optimise

"The last system was built, and nobody uses it."

We build from the top of the roadmap down, one system at a time, each agreed with you in writing before we start. We are full-stack developers first, so it is real software: connected to the systems and data you already have, tested, and monitored once it is live. It is handed over to the same people we interviewed. They asked for the fix, so they are the ones most likely to use it.

// 05

Prove

90 days

Part of Optimise

"How do we know any of it actually worked?"

Ninety days after a system goes live, we measure the hours again against the baseline from step two. Before and after, in the same units, from the same people. If a system has not moved the number, the roadmap says so, and it is fixed or retired before anything else is built.

03What changes

What changes when this works.

The leaks have numbers

You can name the biggest drains on the business and what each one costs a week, instead of a general sense that things are slow.

Work stops happening twice

Information typed into one system and then retyped into another is entered once, and the people who used to do the retyping get those hours back.

The fixes get used

Systems are built for problems the team named themselves, so adoption is not a campaign you have to run afterwards.

Spending follows evidence

Every rand spent on a system traces back to a finding in the report and is checked against the baseline once it is live.

04What it is not

What this is not.

Not a software sale.

Software is one of three answers. When a changed process or stopping a task fixes the problem, that is what the report recommends.

Not a strategy deck.

You get a report with hours and costs in it, a roadmap another developer could build from, and, if you go on, working systems.

Not a performance review.

The interviews are about the work and how it runs. Nothing is attributed to anyone, and recordings never reach management.

Not AI for its own sake.

We use AI where it is the simplest thing that works. Where a form, a spreadsheet or a conversation between two departments does the job, we say so.

Not a long contract.

Each stage stands on its own. The report and roadmap are yours whether or not we build anything.

05How it is bought

Three stages. Stop after any one.

The five steps are sold as three stages, each priced and invoiced on its own.

// 01

The Audit

Listen, Measure, Decide

Every person interviewed, the hours measured, and a report and roadmap you keep.

// 02

The Build

Build

The first system on the roadmap, built and handed over.

// 03

Optimise

Build, Prove

The next systems on the roadmap, and the hours measured again against the audit.

What each stage costs The practical detail of the audit, including POPIA

How efficient is your business running? Your team already knows.

Start with a 15 minute call. There is no charge and nothing to prepare. By the end of it you will know whether the audit is worth your money.

Book a 15-minute call What each stage costs

The system is never the starting point.

So we start there. The interviews find the work. The measurements put a cost on it. Only then do we decide what to build, and only what the evidence supports. After that, we measure again.

Listen to everyone. Measure the cost. Build only what the evidence supports.