✧ Key takeaways

The question is not which AI tool to buy. It is which work to hand over — and which work is the reason clients pay you in the first place.

Here is the question most solo business owners ask first: which AI tool should I buy? It is the wrong first question. The right one is smaller and harder: which piece of my work is repetitive, rule-based, and boring enough that a machine should do it — and which piece is the reason clients pay me in the first place?

You read a thread about someone automating their whole operation, sign up for three tools in one evening, and a week later you are paying for software that sends emails nobody reads while you still copy invoices by hand. The tools were not the problem. The choice of what to hand over was.

This guide fixes the order: a five-question test for any task, the six workflows that pay off first, the work to keep yourself, the real free-tier maths for Zapier and Make, a starter stack that costs nothing, and a 30-day plan. Prices quoted are the published free and entry tiers, checked on 18 November 2026 — check the pricing page before you pay, because those numbers move.

What AI automation for small business actually is — and what it is not

Strip away the marketing and AI automation for a solo business is three ordinary things working together. First, a trigger: something happens — a form is filled in, an email arrives, a payment lands, a calendar slot is booked. Second, a rule: if this happens, do that — file it, reply with a template, create a task, send an invoice, move a card. Third, a small amount of judgement supplied by an AI assistant where a fixed rule is not enough: drafting a first version of a reply, summarising a long thread, pulling the key points out of a call transcript, turning rough notes into a structured brief.

That last part is where people overreach. An AI assistant is good at first drafts, sorting, and summarising text you already have. It is not good at knowing your client, your pricing, or your standards, and it will happily write a confident reply that is wrong in a way only you would catch. So the model is this: automation handles the movement of information; you keep the decisions. A Zap that moves a form submission into your task list is doing its job. A Zap that decides what to charge a client is you handing away the business.

A scheduling tool that lets clients book a slot without emailing back and forth is not really "AI" — it is a calendar with a booking page — but it saves more hours than most AI features will. Judge every tool the same way: how many minutes does it return, how often does it break, and what happens when it gets something wrong?

Automation also has a setup cost, and for a solo operator that cost is your evening. The workflows here are picked because most take under an hour to build and run every day, so the saving compounds.

The 5-question test before automating anything

Run every candidate task through these five questions. If it fails any of the first three, do not automate it yet. If it passes all five, it goes on the list.

1. Does this happen at least three times a week?

Frequency is the whole game. A task you do twice a year is not worth automating, however annoying it is. Write the template, save it in a document, and move on. A task you do three times a week, at five minutes a time, is more than thirteen hours a year. That is the threshold where building a workflow starts to pay. Count the real frequency, not the frequency you imagine. Check your sent mail or your task list for the last two weeks and tally it.

2. Is the rule clear enough to write down in one sentence?

"When a new lead fills in the contact form, add them to the CRM, send the welcome email, and create a follow-up task for tomorrow" is a sentence a no-code tool can follow. "Reply helpfully to client emails" is not. If you cannot state the trigger and the action in one sentence, the task still needs your judgement at the point of decision. You can still automate the parts around it — filing, tagging, drafting — but keep the decision with yourself until the rule gets clearer.

3. What happens when it goes wrong?

Every automation will go wrong eventually. A form field will be blank, an API will time out, a client will write something the template did not expect. Ask what the failure looks like. If the worst case is a task filed in the wrong list, you can live with it. If the worst case is a client receiving the wrong price, a payment reminder sent to someone who already paid, or a personal reply sent to the wrong person, that workflow needs a human check in the middle, or it should not run at all. Build the failure case into the design: send yourself a notification when a run fails, and keep a log you actually look at once a week.

4. Is the input clean enough to trust?

Automation is only as good as the data feeding it. If your inbox is full of half-filled forms, vague subject lines, and clients who reply inside old threads, a workflow that parses those messages will misread them regularly. Fix the input first. Add required fields to the form. Use a booking page instead of open-ended "let me know a time" emails. Give clients a clear subject line to use. Ten minutes spent tightening the input saves hours of fixing bad output.

5. Will I still check it?

Be honest about this one. An automation you never look at is a liability, not an asset. It will keep running after your prices change, after you rename a service, after a client leaves. Pick workflows you are willing to review weekly for the first month — five minutes, checking the log, fixing the one run that failed — and monthly after that. If you know you will never open the log, choose a simpler workflow, or do the task by hand until you have the habit.

Score each task out of five. Five goes first; three or four gets a half-automated version, where the machine drafts and you approve; below three stays manual.

What to automate first: six workflows that pay for themselves

These six are ordered by return, not by excitement. The first three save time every single day. The last three stop things falling through the cracks, which is where solo businesses actually lose money.

1. Inbox triage: sort it before you read it

The inbox is where a solo day goes to die. The fix is not a clever AI that answers everything — it is sorting, labelling, and drafting so you spend your attention only on the messages that need it. A practical setup looks like this: a Zapier or Make scenario watches for new mail, applies a label based on the sender or subject — client, lead, invoice, newsletter, admin — and moves newsletters and receipts out of the main inbox into a folder you read once a day. For messages from known clients, an AI assistant drafts a short reply based on the thread and saves it as a draft, never sending it. You open the inbox, approve or edit the drafts that matter, and delete the rest.

Sorting is rule-based and reliable; drafting is where the AI earns its keep without touching the client. Do not let the tool send replies directly — one bad draft, a wrong date or a promised discount, costs more trust than a month of saved minutes. Keep the send button with you. The realistic saving is 20 to 30 minutes a day once the labels settle.

2. Scheduling: end the back-and-forth

"Does Tuesday work? How about Thursday at 2?" That exchange costs more than the meeting itself. A booking page ends it. A Calendly-type scheduling tool shows your real availability, lets the client pick a slot, sends the confirmation and the reminder, and puts the event on your calendar with the right link. Connect it to a simple follow-up: after a call is booked, create a task in Notion to prepare, and after the call, send a short thank-you note with the next step.

Set a 15-minute buffer between calls and a daily cap so a booking page cannot fill your whole day. Ask one or two questions on the form: what is this about, and is there anything I should read first? Those answers turn a vague call into a prepared one. This rarely needs Zapier or Make; the scheduling tool's own integrations usually cover the calendar, the email, and the task. If you only automate one thing this month, make it this one.

3. Invoicing and follow-ups: get paid without chasing

Most solo businesses do not have a sales problem; they have a follow-up problem. Invoices go out late, reminders never go out, and the money sits in accounts receivable while you do more work. Automate the boring spine of it: when a project is marked complete in your tracker, generate the invoice from a template, send it, and schedule a polite reminder for seven days later if it is still unpaid, and a firmer one at fourteen days. When payment lands, mark it paid, send a receipt, and close the task.

Keep two human checkpoints. First, check the invoice before it goes out — a wrong amount makes a client question everything else. Second, write the reminder in your own voice: "Invoice 104 for the March work, $450, due Friday" gets paid; a system-sounding message gets ignored. The automation handles timing and tracking, not wording. If an invoice is 30 days overdue, no email sequence fixes it — that is a call, and it is yours to make.

4. Social repurposing: one piece of work, three places

Content is where solo operators either burn out or disappear. The sustainable version is repurposing, not creating from scratch every day. Write one solid piece — a short guide, a client story, a breakdown of a problem you solved — and let a workflow turn it into the smaller pieces: a short post for one platform, a few lines for another, a caption with the link. An AI assistant can draft those versions from your original text, keeping your examples and your numbers intact, and you edit before anything is published.

The rule that keeps this honest: the AI drafts from your work; it does not invent new claims. If it adds a statistic you did not write, delete it. If it smooths your wording into generic corporate language, put your wording back. A realistic cadence is one original piece a week and three repurposed posts from it. For the tools behind this workflow, our guide to AI writing assistants compares how they actually work and when each one fits.

5. Lead capture: never lose a form submission again

A lead that fills in your form and hears nothing for two days has already talked to someone else. The fix is a short chain that runs the moment the form is submitted: save the details to your tracker in Notion, send the person a confirmation that says what happens next and when, create a follow-up task for you with a due time, and — if the form asks what service they need — tag it so you can see at a glance what kind of work is coming in. That chain takes about 30 minutes to build in Zapier or Make and runs hundreds of times without you touching it.

Do not add a lead to a marketing list automatically just because they filled in a contact form. Consent for a reply is not consent for a newsletter. Keep the confirmation short: "Thanks — I reply within one working day, and I will send a short quote after a 15-minute call" tells the lead more than a polished welcome sequence.

6. Reporting: a weekly snapshot you actually read

Solo operators rarely lack data; they lack a moment to look at it. Set up a Friday-afternoon summary in one place: new leads, invoices sent and paid, money in, calls booked, content published, and the one task still open from last week. A no-code scenario can gather those from your tracker, invoicing tool, and calendar, and an AI assistant can turn them into three short paragraphs: what happened, what changed, and what needs attention next week.

Keep the report short enough to read in two minutes, and write down the one decision it prompts — "follow up the two unpaid invoices on Monday" is a decision; "revenue is up" is not.

What NOT to automate

The list of things to keep manual is shorter than the list of things to automate, and it matters more. These are the places where your judgement is the product.

First replies to a new client

The first reply sets the tone for the whole relationship. A template that says "thanks for reaching out, I will be in touch" tells a new client nothing about whether you understood their problem. Read the enquiry, answer the actual question, and name the next step. You can use a draft as a starting point, but the reply that goes out should be yours.

Pricing judgement

Never let a workflow quote a price. Pricing depends on scope, urgency, the client's budget, how busy you are, and whether this is work you want. A fixed price list is fine for a standard package; an automated quote for custom work is not. Keep a rate sheet in Notion, but decide the number yourself, every time.

Creative decisions

An AI assistant can suggest ten headlines or ten angles. It cannot tell you which one is right for your audience, because it does not know your audience the way you do. Use it for options and first drafts. Make the choice yourself. The same applies to design, brand voice, and the story you tell about your work — the parts a client cannot get from anyone else.

Anything where a mistake costs trust

Apologies, refunds, scope changes, bad news about a deadline — all of these are human conversations. An automated apology reads as an evasion, even when the words are polite. If something has gone wrong, write the message yourself, say what happened, and say what you will do about it. A client will forgive a late delivery. They will not forgive discovering that a machine apologised for you.

Low-volume tasks

If you do something four times a year, automating it is a hobby, not a business decision. The setup and fixing cost more than doing it by hand four times. A checklist in a document is enough until the frequency rises.

Zapier vs Make: the free-tier maths that decides it

For most solo businesses, the choice between Zapier and Make comes down to the free tier, because that is where you will start — and where you should stay until a workflow proves it earns a paid plan. Both tools do the same broad job: connect apps that do not talk to each other and move data between them on a trigger. They differ in how they count usage and how complex a workflow they allow for free.

Zapier Free costs $0 and includes 100 tasks a month, where a task is one successful action step. Free Zaps are two-step only — one trigger plus one action — with polling every 15 minutes and no premium apps. That is enough for simple chains: a form submission creates a task, a new booking sends an email, a payment marks an invoice paid. The moment you need a second action, a filter, or a faster check, you hit the limit. Zapier Professional starts from $19.99 a month when billed annually, for 750 tasks, so the jump from free to paid is a real step, not a small one.

Make Free also costs $0, but the allowance works differently: 1,000 credits (operations) a month, where each step in a scenario uses credits. The important difference is that the free tier allows multi-step scenarios with routers and filters — so one scenario can check a condition, take one path for new leads and another for existing clients, and update three tools in a single run. Make's paid Core plan starts from about $9 a month when billed annually, for 10,000 credits. Make renamed "operations" to "credits" in 2025; it is the same unit under a new name, so older guides that talk about operations are describing the same allowance.

PlanPriceMonthly allowanceWhat the free tier allows
Zapier Free$0100 tasksTwo-step Zaps only, 15-minute polling, no premium apps
Zapier ProfessionalFrom $19.99/month billed annually750 tasksMulti-step Zaps, faster polling, premium apps
Make Free$01,000 creditsMulti-step scenarios, routers and filters
Make CoreFrom about $9/month billed annually10,000 creditsHigher allowance and more features

Which one should a solo operator pick? If your workflows are simple — one trigger, one action — Zapier's free tier is enough and its app catalogue is the widest. If you want one scenario to do several things or branch on a condition, Make's free tier goes further for $0, and its paid tier is cheaper at entry level. Build your first three workflows in Make Free, see how many credits they actually use, and only then decide whether either paid plan earns its cost. Check the pricing page before you subscribe — the numbers above were verified on 18 November 2026.

One warning about counting: a task or credit is consumed per step, not per run. A five-step scenario running 100 times a month uses 500 credits, not 100. Keep scenarios short.

A starter stack that costs $0

You do not need to buy anything to start. This stack covers the six workflows above for $0, using free plans, and it is the stack to run for the first 30 days before deciding what — if anything — deserves a subscription.

  • Notion as the tracker. Notion has a free plan, and it holds the pieces a solo business needs in one place: a leads table, a client list, a content calendar, and a simple invoice tracker. Keep it simple — four tables, not forty. If you mention Notion to a client or a reader, you can point them to our Notion product page for an overview. The tracker is the hub that your automations read from and write to, so get it right before connecting anything else.
  • Make Free for the connections. 1,000 credits a month covers the six workflows above at solo volume, with room to spare. Use it for lead capture, inbox labelling, reporting, and the follow-up reminders that Zapier's free two-step limit would block.
  • Zapier Free for the simple chains. Keep it for the one-trigger, one-action jobs where its app catalogue is widest. Running both tools is fine at this stage — use each where it is strongest, and consolidate later if one of them earns a paid plan.
  • A ChatGPT- or Claude-type assistant for drafting. Use it to draft replies, summarise threads, and turn notes into structured briefs. Do not state a price for these assistants here — check the pricing page for the plan that fits, because the free and paid tiers differ by model access and limits, and those details change often.
  • A Calendly-type scheduling tool for bookings. The booking page, the reminders, and the calendar sync are the core of it. Check the pricing page for the free tier's limits before you rely on it for paid bookings or team features.
  • Systeme.io for a simple funnel, if you need one. Systeme.io has a free plan, and it covers a landing page, a form, and basic email in one place — useful if you sell a small product or a lead magnet and do not want three separate tools. Check the pricing page before building anything that depends on a paid feature.

Nothing on this stack should cost money in month one. If a workflow needs a paid feature to work at all, it is the wrong first workflow. Our broader setup is covered in The Freelancer's Tech Stack: Every Tool You Actually Need in 2026.

A 30-day rollout plan (week by week)

Thirty days is enough to get four workflows live and to learn which ones you actually use. Do not try to build all six in week one.

Week 1: track and schedule

Set up the Notion tracker — leads, clients, invoices, content — and the booking page. No automation yet, except the scheduling tool's own calendar sync and reminders. Get every lead and invoice into one place, so later automations have clean data. At the end of the week, count enquiries and follow-ups. That is your baseline.

Week 2: lead capture and inbox triage

Build the lead-capture chain first: form to tracker, confirmation email, follow-up task. Test it with a fake submission and fix whatever breaks. Then add inbox labelling and draft replies for known clients only. By Friday, every new lead should be in the tracker within a minute of submitting the form.

Week 3: invoicing follow-ups and reporting

Connect the invoice tracker to the reminder sequence — seven days, then fourteen — and build the Friday report. If the numbers look wrong, the workflow is wrong, not the business. Fix the workflow. Also write down the one manual task that takes the most time, and score it with the five-question test.

Week 4: repurposing and review

Add the content repurposing workflow, using one real piece of your work as the source. Then review all four workflows: open each run log, count the failures, and fix the most common one. Cancel any tool you did not use. Add up the minutes saved in a week. If the total is under an hour, you picked the wrong tasks — adjust before month two, and only then consider a paid plan for the tool that earned it.

Mistakes that waste money

Most of the money solo businesses waste on automation is not spent on the tools themselves. It is spent on the habits around them.

  • Buying before building. Paying for a Professional plan before the free tier has run for a month is guessing. Let the free tier tell you the real usage.
  • Automating a broken process. If your intake form asks the wrong questions, automating it just delivers bad data faster. Fix the process on paper first.
  • Too many tools doing the same job. Three automation tools, two schedulers, and four note apps is not a stack; it is a subscription pile. Pick one tool per job and delete the rest.
  • No owner for the workflow. In a one-person business, you are the owner. That means a monthly five-minute check: open the log, look at the failures, update the template when a price or a service changes. A workflow with no owner rots quietly.
  • Automating the reply, not the task. The most expensive mistake on this list. A client who receives a wrong automated reply does not think "the tool made a mistake". They think you did not care enough to write it yourself. Draft with AI, send as yourself.
  • Counting the saving wrong. A workflow that saves 10 minutes a week but takes an hour a month to maintain saves nothing. Count the maintenance, not just the headline saving, and retire workflows that cost more than they return.

If you recognise two or more of these, do not add another tool. Spend a week fixing the ones you have.

Bottom line: automate the movement, keep the decisions

The real question was never which tool to buy. It is which work is repetitive enough to hand over, and which work is the reason clients pay you. Run the five-question test, start with scheduling and lead capture, keep first replies and pricing with yourself, and let the free tiers prove their value before you pay. Zapier Free covers simple chains at 100 tasks a month; Make Free covers multi-step work at 1,000 credits a month.

Get the automation starter checklist: write down the six workflows from this guide, score each task with the five-question test, and build one per week for four weeks — scheduling and tracking, then lead capture and inbox triage, then invoicing follow-ups and the Friday report, then content repurposing. At the end of the month, add up the minutes saved, cancel the tools you did not use, and keep only the workflows that earned their place.

How we checked this: research-based guide compiled from the official Zapier, Make, and Notion pricing pages and independent 2026 comparisons, verified 18 November 2026. No hands-on testing claim is made for any tool in this guide, and prices should be checked on the official pricing page before purchase, as plans and allowances change.

Featured in this guide

Mors

Mors is a senior technology analyst and developer who builds, tests, and breaks modern software platforms to find what truly works. Specializing in cloud engineering, optimization tools, and SaaS architecture, they provide data-driven, objective reviews designed to save development teams time, money, and computing overhead.