13 November 2026
5 Automations a Small Business Can Set Up in a Week

“Automation” sounds like a big project, but some of the most useful workflows are small. They take a task someone does by hand, several times a day, and make it happen on its own. Here are five that suit a small business, what each one needs, and what can go wrong. None needs custom software.
What “in a week” really means
A simple workflow can be built and tested in a few days. What takes the time is usually not the building:
- Decisions: who gets which enquiry, what the reply says, when a reminder goes out
- Access: logins to your forms, ad accounts, calendar, sheet or CRM
- Approvals: sign-off on wording, especially anything that messages customers
- Testing: running it with real examples before you trust it
If those are ready on day one, a week is realistic for one workflow. If they aren’t, the week stretches. Start with one automation, not five.
Automation 1: Send new enquiries to the right person immediately
The problem: a form is submitted at 9:40pm and sits in an inbox until tomorrow. By then the customer has contacted a competitor.
What it does: when a form is submitted, or a Facebook lead form is filled in, the details are captured, tidied, logged in one place, and sent to the right person as a WhatsApp or Telegram alert.
What you need: your form or lead source, one place to log leads (a shared sheet is fine), a notification channel your team watches, and a simple rule for who gets what.
Watch out for: missing phone numbers, duplicates, and what happens when the person is away. Our posts on automating lead routing and on webhooks explain how this works.
Automation 2: Reply instantly, even out of hours
The problem: customers who write in the evening hear nothing until the morning, and wonder if anyone received their message.
What it does: the customer gets an immediate acknowledgement saying their message arrived and when to expect a reply. For simple factual questions, such as opening hours or location, it can send an approved answer.
What you need: approved wording, your working hours, and a decision on what counts as a simple question.
Watch out for: don’t let it answer anything that needs judgment, and always give a way to reach a person. If messages go through the WhatsApp Business Platform, there are rules on message types and templates. Our comparison of chatbots and human support covers where to draw the line.
Automation 3: Reminders that people actually see
The problem: appointments get missed, and invoices get forgotten, and staff spend time chasing.
What it does: at a set time before an appointment, or a set number of days after an invoice is issued, a reminder goes out automatically. If someone confirms or pays, the reminders stop.
What you need: a source of dates (a calendar or booking tool), the reminder wording, and a way to see whether the person responded.
Watch out for: reminders sent after the customer has already paid or cancelled. That’s worse than no reminder, so build in the stop condition and test it.
Automation 4: A weekly report that arrives on its own
The problem: someone spends an hour or two every week copying numbers from ads, analytics and a sheet into a report.
What it does: at a fixed time each week, it pulls the numbers you care about from your sources, puts them in one summary, and sends it to you.
What you need: a short list of the numbers that matter (leads, cost per lead, enquiries by source), access to each source, and one place to send it.
Watch out for: automating a report nobody reads. Start with a small set of numbers, and make sure they match your own records. Our posts on what a good cost per lead is and how to tell if your ad agency is working cover which numbers deserve a place.
Automation 5: Ask for a review after a good experience
The problem: happy customers rarely leave reviews unless asked, and asking by hand gets forgotten.
What it does: a few days after a purchase or visit, it sends a polite message with a link to leave a review, and nothing more.
What you need: a trigger (a completed order or appointment), the right review link for your business, and timing that suits your customers.
Watch out for: only ask customers who had a good experience, and don’t offer incentives that break a review platform’s rules. Reviews also help your visibility in local and AI search, which we cover in our guide to making your website AI-ready.
Which one to start with
| If your biggest problem is… | Start with |
|---|---|
| Leads going cold or being answered late | Automation 1, then 2 |
| Customers asking the same questions at all hours | Automation 2 |
| No-shows or late payments | Automation 3 |
| Hours lost on reporting | Automation 4 |
| Few reviews or little social proof | Automation 5 |
Pick the one that costs you the most, measured in lost enquiries or wasted hours. Our post on whether AI automation is worth it shows how to work that out.
What you need before you start
- One task, described step by step in plain language
- The decision rules written down (who, when, what)
- Logins for every system the workflow touches
- Approved wording for anything customers will see
- Someone who will own it afterward
- A way to measure it: time to first reply, reminders sent, reports delivered
A one-week plan
| Day | What happens |
|---|---|
| 1 | Pick one automation. Write the steps and the rules. Gather access. |
| 2 | Agree wording and decide what “done” looks like. |
| 3 | Build the workflow with the simplest version that works. |
| 4 | Test with real examples, including awkward ones: missing details, duplicates, odd hours. |
| 5 | Add alerts for failures and a manual fallback. Fix what testing found. |
| 6 to 7 | Run it alongside your normal process. Compare results, then switch over. |
What can go wrong
- It fails silently. A workflow that stops without anyone noticing is worse than the manual process. Set up an alert for failures, and check it periodically.
- Wrong or unapproved messages go out. Review wording first, and test with your own number.
- Personal data ends up in the wrong place. Automations move customer details between systems. Know where the data goes, who can see it, and limit it to what’s needed. In Malaysia, the Personal Data Protection Act applies.
- Nobody owns it. When a tool changes or a login expires, the workflow breaks. Name an owner, and write down how it works.
- Over-automating. Customers still want a person for complaints, exceptions and big decisions.
What not to automate yet
- Tasks that need real judgment each time
- Processes that differ every time, or that nobody has written down
- Anything where a mistake would harm a customer, such as health or money decisions, without a person reviewing it
- Tasks that happen so rarely that building the workflow costs more than doing it by hand
How to know it worked
Compare a few numbers before and after, over the same period:
- Time from enquiry to first reply
- Enquiries answered within your target time
- Hours spent on the task each week
- No-shows, late payments or reviews collected, depending on the automation
- Mistakes or complaints caused by the workflow
If the numbers haven’t improved, adjust or stop. A small test costs little, and the result tells you whether to do more.
Want one of these built, tested and monitored for your business, with the rules agreed before anything goes live? [See how we build automation workflows →]