Automation
7 Repetitive Tasks Small Businesses Can Automate
Seven concrete, low-risk tasks in service businesses that are usually worth automating first, and what each one needs to work properly.
Lucinda de Bruin · 2 July 2026 · 4 min read
"Automate your admin" is easy advice and difficult to act on. So here are seven specific tasks that come up again and again in South African service businesses, what each one actually needs, and where a person should stay involved.
None of these require a technical team. All of them require agreement on how the work is currently done.
1. Enquiry acknowledgement and triage
The task: A new enquiry arrives by email, form or WhatsApp. Someone reads it, works out what it is, and eventually replies.
What automation looks like: The enquiry is captured in one place, summarised into a consistent short brief, categorised, and acknowledged immediately. Anything urgent or unusual is flagged for a person.
What it needs: A single intake point, and a written list of your categories.
Where a person stays: Deciding whether to take the work, and any pricing conversation.
2. Quote and proposal first drafts
The task: Preparing a quote from scratch each time, hunting for the last similar one.
What automation looks like: A draft assembled from your own templates, current rates and the details of the enquiry, ready for you to adjust and approve.
What it needs: Your pricing rules written down, and a few good past examples.
Where a person stays: Final pricing, scope decisions, and sending. Draft, never auto-send.
3. Follow-up on open quotes
The task: Chasing quotes you have already sent. Usually the first thing to slip in a busy week.
What automation looks like: A defined sequence (a check-in after a few days, a second after a week or two) drafted in your tone, with the whole sequence stopping the moment the client replies.
What it needs: A reliable record of which quotes are open. Usually a simple pipeline.
Where a person stays: Approving the sequence, and any negotiation.
4. Meeting notes into actions
The task: A useful call happens, and three days later nobody is sure what was agreed.
What automation looks like: The recording or your notes are turned into a short summary, a list of agreed actions with owners, and a draft follow-up message.
What it needs: Consent to record where you record, and a place where actions live.
Where a person stays: Confirming the actions are correct before they are shared.
5. Client onboarding admin
The task: The same twelve small steps every time a client says yes.
What automation looks like: One trigger creates the folder structure, the project record, the task list, the intake request and the welcome message.
What it needs: An honest list of every step, including the ones nobody has written down.
Where a person stays: The welcome itself, and anything specific to that client.
6. Recurring reports
The task: Rebuilding the same weekly or monthly report by copying numbers between tools.
What automation looks like: Figures pulled from the systems you already use on a schedule, assembled into the same format, with a short written summary of what changed.
What it needs: Agreement on which numbers matter and where each one comes from.
Where a person stays: Interpretation. A system can tell you the number moved; you decide what it means.
7. Answering internal "where is that" questions
The task: Staff asking each other for the policy, the current rate card, the process for a specific situation.
What automation looks like: An assistant that answers questions against your own documents and links to the source, so answers can be checked.
What it needs: Documents that are current. This one exposes out-of-date material quickly, which is itself useful.
Where a person stays: Owning and updating the underlying documents.
Choosing between them
If several of these apply to you, choose using two questions rather than enthusiasm:
- Which one is costing the most time each week?
- Which one is the most agreed-upon internally?
Where those two overlap is where to begin. If nothing overlaps, start with the most agreed-upon one anyway: the value of a first success that actually gets used is higher than the value of the theoretically largest saving.
A note on expectations
Each of these removes handling time and reduces the number of things depending on someone's memory. None of them removes the need for judgment, and none of them is finished on the day it goes live. Plan for a review a few weeks in, once real work has passed through it.
If you would like help deciding which of these fits your business, the AI Opportunity Audit is built for exactly that decision, and How It Works sets out the process we follow.