How do you tell whether a task can be automated?
Process automation works when a task can be described like a recipe: “when X happens, do Y”. If you can explain it to a colleague in three sentences without a single “it depends”, it is a good candidate.
Five traits, taken together, point to a task worth automating: it is repetitive, it follows fixed rules, it is frequent, it uses data that is already digital (an Excel sheet, a management system, an email inbox) and it is easy to get wrong by hand, through distraction or fatigue.
8 repetitive tasks to automate in your business
These are the automations that businesses and professionals ask for most often. For each one you will find what it does, who it is for and an example we built to show how it works.
1. Reminders for unpaid invoices
The system checks due dates and sends automatic payment reminders by email and WhatsApp at 7, 14 and 30 days, with a firmer tone each time, and logs the payments received. It suits anyone who issues many invoices and hates chasing clients: professional firms, suppliers, tradespeople. You can try it in the invoice reminder demo.
2. Appointment reminders
The client receives automatic reminders by email or SMS 24 hours and 2 hours before, and can confirm or reschedule with one click. You get a daily summary. It is designed for medical practices, salons, beauty centres and consultants. Here is the appointment reminder demo.
3. Daily sales report
An automatic report gathers sales from the shop, the website and phone orders and emails you a PDF with the day's numbers every evening. It suits anyone selling on several channels who copies the data into Excel by hand today. See the sales report demo.
4. Low-stock alerts
When a product drops below the threshold you set, you get an alert before you run out. It suits shops, restaurants, workshops and anyone with a stockroom, however small. Try the inventory alert demo.
5. Social post scheduling
You prepare the week's content in one session and the system publishes it on the days and at the times you set. It suits anyone who runs their own profiles and ends up posting “whenever there's time”. There is a social scheduling demo.
6. Onboarding emails for new clients
As soon as a client signs or buys, a sequence of emails starts: welcome, documents to send, next steps, useful contacts. It suits firms, agencies and subscription services that explain the same things every time. See the onboarding sequence example.
7. Review requests after a service
A few hours or days after the visit, the client receives a message with a direct link to leave a review on your Google Business Profile. It suits restaurants, practices, garages and any local business that lives on word of mouth. Here is the review request example.
8. Contracts and quotes generated from data
You enter the client's details and the chosen items once and the system fills in the document on your template, ready to send. It suits anyone who copies and pastes from an old quote today, at the risk of leaving the previous client's name in. Try the quote tool demo or look at the contract generator example.
How much time does automation save?
It depends entirely on how much time you spend on it today, so you can do the maths yourself. Take one task at a time and multiply the time it takes by the number of times you do it.
| If today… | Your maths |
|---|---|
| you spend 2 hours every Monday checking who hasn't paid | 2 hours × about 48 weeks = about 96 hours a year |
| you spend 20 minutes every evening adding up sales | 20 minutes × 6 days × 48 weeks = 96 hours a year |
| you make ten 2-minute calls every day to remind clients of appointments | 20 minutes a day, which add up as above |
On top of those hours, add what you don't see: the forgotten invoice, the appointment missed without notice, the product that runs out on the busiest day. These are costs only you know, and they weigh as much as the time.
What should you not automate?
Automating everything is not the goal. There are three kinds of task that are better left to a person.
- Decisions that need judgement. Granting a discount, assessing a complaint, choosing a supplier: the system can prepare the data, but the choice is yours.
- Moments that matter for the relationship. An angry client, an important negotiation, a serious delay to announce: an automatic message makes things worse here. Even the third reminder to a long-standing client deserves a phone call.
- Processes that change every month. If the rules are not stable yet, automating them means redoing the work at every change. Sort them out first, then automate.
How an automation project works
The method we follow is designed to show a concrete result early and to spend only where it is needed.
- Map the process. Together we write down every step as you do it today: where the data comes from, who handles it, where it ends up. This often reveals pointless steps to remove before automating anything.
- Start with a single automation. We pick the one with the best ratio between hassle removed and complexity. It usually connects to the programs you already use, through an API or a direct integration, without changing the way you work.
- Test with real data. For a few weeks the automation runs while you check the results. Edge cases get fixed: the client with two email addresses, the invoice paid in two instalments.
- Measure the time you get back. You do the maths, with the same questions as the table above: how long it took before, how long it takes now.
- Extend. Only when the first automation works well do we move on to the second, often reusing the same connections.
Technically, automations start in one of two ways: at a fixed time, with a cron job that for example generates the report every evening, or when something happens, with a webhook that reacts to a new order or a payment. You will find all the demos to try on the services page.
Mistakes to avoid when automating
- Automating a messy process. If today the data is spread across three different sheets and nobody knows which one is current, the automation will just produce errors faster. Tidy up first.
- Trying to do everything at once. A project touching ten processes at the same time takes months before it delivers anything. A single well-built automation shows results within weeks.
- Messages that sound like a robot wrote them. Reminders should be written in your tone, signed by you, with an easy way to reply to a person.
- Nobody keeping an eye on it. Every automation must alert you when something goes wrong: an undelivered email, a missing piece of data. A system that fails silently is worse than doing it by hand.
- Forgetting about privacy. Automations process client data: you need the right legal basis and attention to the GDPR, especially for messages that are not about a service already purchased.
Frequently asked questions
01Do you need custom software to automate business processes?+
Not always. Many automations are built by connecting the programs you already use, such as your management system, email and Excel sheets. Custom software is needed when the data does not exist in digital form or when no existing tool follows your rules.
02How long does it take to set up an automation?+
It depends on the complexity and on how many programs need to be connected. A single automation on tidy data, such as a reminder or a payment chaser, can be set up quickly; projects touching several processes take longer and are best split into phases.
03Can an automation make mistakes?+
Yes, if the source data is wrong or if an unforeseen case comes up. That is why it is tested with real data before being left to run on its own, and why it includes an alert when something goes wrong.
04Which automation should you start with?+
With the one that costs you the most time or money when you forget it. For many businesses that means invoice reminders or appointment reminders: simple rules, data already available, an effect you can see straight away.
05Do I have to change the programs I already use?+
Usually not. The automation builds on the tools you have, as long as they allow the data to be read and written. If a program is too closed, we look at the simplest alternative together.
