Avodah means work, worship, and service in the same breath. It is the word our institute is named for, and it carries a claim most business training never makes: that the work itself is accountable to something beyond its output.
Which raises a question our founders rarely get asked before they are sold a tool.
If work is worship, what is a step that serves no one?
The rush that is costing African founders money
MIT’s GenAI Divide study examined 300 enterprise AI deployments and found that 95 percent produced no measurable profit impact, against $30 to $40 billion of spend. Those were companies with budgets, consultants, and time to waste.
Our founders have none of those. A founder in Kampala paying for three AI subscriptions on a business turning over four million shillings a month is not running a pilot. They are running an experiment they cannot afford to lose.
The pressure is real and it is not imaginary. Everyone in the ecosystem is being told that AI is the difference between surviving and being left behind. The fear is doing the buying.
Two questions, asked in order
There is a framework from AI consulting practice, taught as QDOAA, that insists automation is the last of five moves rather than the first.
| Step | The question | The stewardship reading |
|---|---|---|
| Question | Why does this step exist? | Can you name who this serves? |
| Delete | What can go entirely? | Work that serves no one is not work, it is habit |
| Optimize | How do we do this better by hand? | Faithfulness in small things, before scale |
| Accelerate | How do we make it faster without more people? | Good stewardship of the team you already have |
| Automate | Now, where does AI genuinely fit? | Tools serve the work, never the reverse |
Underneath it sits a second question, sharper and more personal. It comes from a filter designed for exactly the overwhelm our cohorts describe: why are you learning this, or building this, or buying this? To what end? If you cannot answer clearly, you are moving from fear rather than from fact.
Fear-based action and fact-based action look identical from the outside. Both involve a founder working late. Only one of them compounds.
What this looked like for one founder
A cohort member ran a small catering business supplying corporate events in Kampala. She came to a session wanting help building an AI system to write her quotations, because quoting was eating her evenings.
We walked the process instead of building the tool.
Seven steps. A client enquiry arrives on WhatsApp. She asks for headcount. She asks for date. She asks for dietary requirements. She checks supplier prices. She writes the quote in a Word document. She converts it to PDF. She sends it back on WhatsApp.
Question: why are the first three steps separate messages? Because that is how the conversation naturally goes. Delete: nothing yet. Optimize: one message asking all three questions at once, saved as a WhatsApp quick reply. Accelerate: a spreadsheet with supplier prices already loaded, so step four becomes a lookup rather than five phone calls.
Seven steps became four. Her evenings came back. Total spend: nothing.
The AI quotation system she came for was never built, because after the optimisation there was not enough left to justify it. She spends the money on a delivery rider instead, which is throughput she can see.
That is not an argument against AI. It is an argument for earning the right to use it. When her volume triples, she will automate, and she will automate four clean steps rather than seven muddled ones. It will cost her less and break less often.
The discipline this asks of us
There is a temptation, in an ecosystem hungry for relevance, to teach the tools because the tools are what people came for. Our founders arrive asking which AI to use. Answering that question directly feels like service. It is usually not.
The AVODA position holds: AI is only useful when it is governed by truth, judgment, service, and accountability. Truth means the process map matches what actually happens, not what the org chart claims. Judgment means knowing which of the five letters you are actually on. Service means the automation serves the customer, not the founder’s sense of being modern. Accountability means someone owns the system after the consultant leaves.
Strip any one of those out and you get the 95 percent.
The triple bottom line, applied
Run the QDOAA pass and the returns arrive on all three lines at once.
Revenue. Fewer steps means lower cost to serve, faster quotes, and money not spent on subscriptions that automate waste. The founder above added a rider instead of a tool.
Social impact. Deleting the step that adds nothing frees a person for work that does. In a business with four staff, that is not a productivity statistic. That is one person’s day.
Kingdom value. Work that serves no one is not neutral. It consumes hours that belong to a family, a church, a community. Questioning it is not efficiency. It is stewardship of the only non-renewable thing a founder has.
Our vision has always been a thriving Christian ecosystem where believers build confidently with and within Africa. Confidence is not the same as speed. A founder who has questioned every step of their process and can defend each one is more confident than a founder with a stack of subscriptions and a vague hope.
For the cohort
Before the next session, take one process in your business. Write down every step, including the ones that feel too obvious to write.
Against each step, answer one question: who does this serve?
Bring the list. We will run the other four letters together.
Do not buy anything first.