Small Business Automation: Where to Start

Small business automation means letting software do the repeatable parts of your week: copying a web lead into the CRM, sending the confirmation and the reminder, chasing an unpaid invoice, building the Monday numbers. Start with lead capture and follow-up, because a slow reply costs you the sale. Check what the tools you already pay for can do before you buy anything. When you need to connect two apps that don't talk to each other, Make, Zapier and n8n all have free tiers and paid plans from about $12 to $50 a month, checked on their pricing pages in October 2026. The rest of this post is how I'd pick the first project and what each tool actually costs.

What counts as automation for a small business

Most of it is rule-based: when this happens, do that. A form gets submitted, so a contact is created, an email goes out and someone gets a task. Nothing in that chain needs judgment, and that's why it's safe to hand off.

It comes in two flavors. Built-in automation lives inside one app, like an invoicing tool that sends its own payment reminders. Connector automation moves data between apps, and that's where Make, Zapier and n8n come in.

AI changes one thing. Steps that used to need a person reading something, like sorting an email into billing or support, or pulling the job address out of a free-text message, can now be done by a model inside the same workflow. I'll get to where that's worth it at the end.

Five processes worth automating first

These are the ones I'd look at in almost any small business, roughly in order.

Lead capture and follow-up

A lead fills out your form. Without automation, somebody notices the email, copies the details into the CRM, writes a reply and sets a reminder to follow up. With it, the form submission creates the contact, sends a reply within a minute, alerts whoever owns sales and schedules the follow-up tasks.

This goes first because speed matters here more than anywhere else, and because it's the process most likely to be done inconsistently by hand. We wrote a full guide to automating lead follow-up.

One thing I'd copy from how we built our own: save the lead somewhere durable before anything else runs. Our website's lead webhook writes every submission to a Postgres database first and only then notifies n8n, which handles the CRM sync. If n8n is down or a step fails, the lead is still there and the failed delivery is logged so it can be re-sent. Even if all you have is a Google Sheet, make that the first step.

If your follow-up includes texts, the rules apply before the automation does. Marketing texts need prior express written consent, and A2P 10DLC registration has been required for business texting from local numbers since February 1, 2025. Registering is not consent. Our SMS compliance guide covers both. This is not legal advice.

Scheduling and reminders

If people book time with you, a scheduling tool with automatic reminders pays for itself in fewer no-shows. Calendly Workflows, for example, can send reminder emails and texts before a meeting without any connector tool. Add a step that creates the client record and the prep task when a booking comes in, and nobody has to touch the calendar.

Invoicing and payment reminders

Invoices get sent late because they're boring, and reminders get skipped because they're awkward. Both are easy to automate. QuickBooks Online has automatic invoice reminders built in: you can schedule up to three, from 90 days before to 90 days after the due date, and it only sends them for unpaid invoices you've already emailed. A connector tool adds the part before that, creating the invoice when a job is marked done in your project tool.

Client onboarding

When someone pays, the same five things usually happen: a welcome email, a CRM update, an intake form, a kickoff meeting link and a set of internal tasks. That's a checklist, and a checklist is a workflow. Trigger it from the payment in Stripe or the signed proposal, and every new client gets the same start whether you're busy or not.

Weekly reporting

A scheduled workflow can pull last week's revenue from your payment processor, new leads from the CRM and open tasks from your project tool, then post a summary to Slack or email every Monday morning. It's rarely the biggest time saver, but it's the one people notice, because the numbers show up without anyone asking for them.

Check the tools you already pay for

Before you add a connector, look at the settings of what you have. Invoicing tools send reminders. Scheduling tools send confirmations. Most CRMs and email platforms have their own workflow builders for follow-up sequences. These built-in automations are free with your plan, they break less because there's no second system involved, and the vendor supports them.

You need a connector tool when the trigger is in one app and the action is in another: a form in one tool, a contact in a different CRM, a message in Slack.

What Make, Zapier and n8n cost

All three have free tiers. They count usage differently, and that matters more than the sticker price.

Prices from each vendor's pricing page, October 2026.

The counting rules change the math. In Make, an operation is a single module run that processes or checks for data, and by default each one costs a credit. A scheduled trigger that checks for new rows and finds none still runs a module. AI modules are billed by tokens instead. In Zapier, triggers and filters don't count as tasks, only the action steps that complete. In n8n, a run is one execution no matter how many steps it has, so a 12-step workflow costs the same as a 2-step one.

My short version: Zapier is the easiest to start with and connects to the most apps. Make is cheaper for workflows with many steps and gives you more control over branching. n8n is the best value once workflows get long or run often, and the Community edition is free if you're willing to host it. The trade-offs are covered in our Make vs n8n vs Claude Code comparison, and if you want to try n8n, start with the n8n beginner tutorial.

How to pick your first project

Spend a week writing down every task you or your team repeat. For each one, note how often it happens, how long it takes and what goes wrong when it's done by hand. Then pick the task that is frequent, rules-based and annoying, with a clear trigger you can point at.

A clear trigger matters more than people expect. "When the form is submitted" is a trigger. "When a client seems unhappy" isn't, at least not until you add AI.

Here's a hypothetical to show the math. Say someone on your team spends 20 minutes on each new lead, between data entry, the first reply and setting up follow-up tasks, and you get 30 leads a month. That's 10 hours a month. If automation cuts it to 3 minutes a lead for review, you get back about 8.5 hours. At $30 an hour that's around $255 a month, against a connector plan that costs $12 to $50. Do the same math on your own numbers, and count the hours to build and maintain the workflow, since that's where people underestimate. Our workflow automation ROI calculator walks through it.

Then build it in this order:

1. Write the current process down step by step, including who does each step.
2. Build the trigger and the first action only, and test it with fake data.
3. Add the remaining steps one at a time, testing after each.
4. Decide what happens when a step fails. Most tools can email you on an error. Turn that on before you turn the workflow on.
5. Watch the run history for the first two weeks.

Where AI agents fit

Rules handle the steps where the input is predictable. AI is useful for the steps where a person used to have to read something first: classifying an inbound email, pulling details out of a messy message, drafting a reply for someone to approve. Make, Zapier and n8n all have AI steps you can drop into a workflow for this.

An AI agent goes further. It decides which steps to take instead of following a fixed path, and it can hold a conversation. The chat assistant on our own website is one. It answers questions from a knowledge base and collects contact details. Every conversation is stored in the same Postgres database our forms write to, and it becomes a lead only once it includes an email or phone number.

I'd still start with rules. A rule-based workflow does the same thing every time and is easy to debug. Add AI to the one step that needs judgment, and keep a person approving anything that goes to a customer until you've read enough of its output to trust it.

If you'd rather have the first workflow scoped and built for you, AI consulting for small businesses takes one process from idea to working in weeks.