Every week someone forwards me a list of "50 AI tools that will transform your business." I have stopped reading them. Not because the tools are bad, but because a list of fifty tools is not advice — it is a way of avoiding the only question that matters, which is: which two or three are worth paying for, for a real business, this year?
This essay is my answer. It is short and opinionated on purpose. I am writing it for a Tanzanian small or medium business — a shop, an agency, a school, a clinic — earning in shillings and watching every recurring cost. If you run a venture-funded startup with dollars to burn, this is not for you.
The rule before the list
Before any specific tool, one rule that will save you more money than the rest of this essay: pay for AI the way you pay for a new employee, not the way you pay for electricity.
An employee has to justify their salary with output you can point to. Electricity you just pay, forever, without asking. Too many businesses sign up for AI subscriptions and then treat them like electricity — a cost that simply exists. Every AI tool you pay for should have a job, and you should be able to name what it replaced. If you cannot, you are not adopting AI. You are donating to a software company.
Pay for an AI tool the way you pay an employee, not the way you pay for electricity. If you cannot name the job it does, cancel it.
1. A general assistant — worth it for almost everyone
The single highest-value AI purchase for most small businesses is one good general-purpose assistant — the kind you chat with. One paid seat, used seriously, replaces a surprising amount of work: drafting quotes, rewriting supplier emails in cleaner English, summarising long documents, turning your rough notes into a presentable proposal.
The reason to pay rather than use the free version is not magic features. It is reliability and the better model — the paid tier makes fewer mistakes on exactly the work you are trusting it with. One seat is enough to start. Do not buy seats for the whole team until one person has proven, in writing, that it earns its keep.
What it replaces: several hours a week of writing and reading you currently do yourself or pay someone to do.
2. A transcription and meeting tool — worth it if you talk for a living
If your business runs on conversations — sales calls, client meetings, site visits — a tool that records, transcribes, and summarises them is quietly transformative. The win is not the transcript. It is never again losing the detail of what a client agreed to because nobody wrote it down.
A caution specific to here: test it on Kiswahili and on Tanzanian-accented English before you pay. Many transcription tools are excellent in American English and useless on the way we actually speak. Run a real meeting through the free trial first. If it cannot handle how your clients talk, no monthly fee will fix that.
What it replaces: the notes nobody took, and the disputes that follow.
3. One automation for your single most repetitive task — worth building, not buying
The third one is not a product you subscribe to. It is a small AI-powered automation built around the one task that wastes the most of your week — answering the same WhatsApp questions, sorting incoming enquiries, generating the same report every Friday.
This costs a developer-week to build, roughly, and then runs cheaply. The reason I list it as worth paying for is that it is the one AI investment shaped entirely around your business rather than a generic tool's idea of business. Build it once, for the task you can name in your sleep, and it pays for itself far faster than any subscription.
What it replaces: the specific repetitive task you hate most. You already know which one.
What I deliberately left off
No image generators, unless design is your business. No "AI CRM," "AI analytics," or "AI everything" suites — they bundle one useful feature with ten you will never open, and charge for all eleven. No tool whose pricing is in dollars and whose value to you is marginal, because a weak shilling will turn "marginal" into "not worth it" in a single bad month.
The absences are the point. A focused business does not need fifty tools. It needs a general assistant, maybe a transcription tool, and one custom automation around its worst repetitive task. Three things, each with a nameable job, each earning its keep.
What to do this week
Do not buy anything yet. Instead, write down the three tasks that wasted the most of your time last week. Then ask, for each, whether one of the three categories above would genuinely have helped — and by how much. Buy the one with the clearest answer, on a monthly plan, for one person. Give it a month. If you cannot point to the hours or the shillings it saved, cancel it without guilt and try the next.
That discipline — adopt one, measure, keep or cut — is the entire skill of buying AI well. The tools will keep changing. The discipline does not.