"Automation" sounds like a big investment in robots and AI. For a small business it's usually much smaller and much more useful than that: stop repeating a clear, rules-based admin step while keeping a person responsible for decisions and exceptions.
What's actually worth automating
The honest answer is: anything you do repeatedly that follows the same steps. The frequent culprits we see in small businesses:
- Quote generation. Same template, same calculation, different customer. A short form on the website plus a templated document can prepare a draft for human review.
- Booking and reminders. Calendly, Cal.com or a booking widget on your site removes the back-and-forth and the no-shows.
- Follow-ups. "Did you have any more questions about the quote?" three days later can be scheduled, but only if a reply, decline or acceptance cancels it.
- Data entry between tools. New customer in your CRM → row added to a spreadsheet, email sent, invoice template created. This is the bread and butter of Make/Zapier-style tools.
- Reporting. If you copy numbers from one place to another every Monday morning, that's a script.
- Social posting and review requests. Both are improved by being scheduled consistently rather than done in bursts.
What's not worth automating (yet)
It is genuinely possible to over-automate a small business. The signs:
- The task happens twice a year. Spending three days automating it saves you five minutes.
- The "rules" change constantly. Automation rewards consistent patterns; if every case is different, a human is still cheaper.
- The customer experience is worse for it. Auto-replies that sound robotic at first contact lose more business than they save in admin.
The test is simple: are you saving time without losing quality, or are you saving time and quietly making the experience worse?
The tools that genuinely matter
| Task | Possible tools | Approach |
|---|---|---|
| Routine form, sheet and email work | Google Forms / Sheets / Gmail | Built-in features and configuration |
| Connect services in a visual workflow | Make / Zapier | Visual no-code workflow editors |
| Self-hosted or more complex branching | n8n | Visual workflow editor; setup and operation may need technical knowledge |
| Custom logic inside Google services | Google Apps Script | JavaScript in a browser-based code editor |
| Small operational data sets | Airtable / NocoDB | Table-based databases |
| Scheduling | Calendly / Cal.com | Booking services |
Make and Zapier are visual no-code workflow editors. n8n also has a visual editor, although running it can require more technical knowledge. Google Apps Script is different: Google documents it as a browser-based JavaScript code editor that runs scripts on Google's servers. Choose by task, access, support and data requirements—not by tool popularity.
Off-the-shelf vs custom
Check an existing service or a small visual workflow first. It is often quicker to test and easier to support. The right answer depends on the process; there is no useful universal percentage.
Custom-built tools start making sense when:
- You're paying per-seat for software you only use lightly, for ten team members.
- The off-the-shelf tool does 70% of what you need and the missing 30% is the whole point.
- Your workflow is your competitive advantage and a generic tool flattens it.
- The data is sensitive enough that you don't want it on a third-party platform.
That's the territory we cover with custom digital tools.
How to actually start (this week)
- List your week. Write down every recurring task you do for one week. Don't edit. Just write.
- Tag each one. Could a clear set of rules do this? Yes / No / Partly.
- Pick one Yes that you do at least twice a week. That's your first project. Don't pick the most complex; pick the most frequent.
- Spend two hours building the smallest possible version. A Make scenario, a Sheet, a Calendly link, a Gmail template. Don't optimise yet.
- Use it for a fortnight, then improve it. Real use shows you what's wrong faster than planning does.
A concrete example
- Enquiry formCapture the required details
- Check scope and ratesDetect missing details or exceptionsException routeIf details are missing or out of scope, stop for review
- Draft quoteFill the agreed template
- Human review and approvalJudge the price and termsHuman decision
- SendOnly send approved content
- Conditional follow-upOnly when no reply is recorded
Illustrative estimate: if a business prepares 10 quotes a week, and each currently takes 20 minutes, that is 200 minutes. If automation prepares the information and draft while a person spends 5 minutes reviewing each quote, review takes 50 minutes. The illustrative saving is 150 minutes a week before setup, maintenance and exception handling. Measure your own volume and review time during the fortnight trial rather than treating this example as a promise.
Where to go next
Read our related article on 5 admin tasks every small business should automate for five copy-and-paste-style projects. And if you'd rather have someone build a couple of these for you properly, our workflow automation service covers connected processes; where the work needs a dedicated interface, see custom digital tools.