A Lean AI Stack for a Small Business (Under $300 a Month)

Most of the AI spend I see in small businesses is not too large. It is scattered. There is a writing assistant somebody signed up for in February, a chatbot on a plan sized for ten times the traffic, a scheduling tool nobody has opened since the person who bought it changed roles, and a separate transcription service doing work the phone system already does.

Organising AI tools for small business by the job they do, rather than by the brand on the invoice, fixes most of that. There are five jobs in a service business, and once you name them you can see which subscriptions overlap and which job has nothing pointed at it. A useful stack for a business with two to fifty staff fits comfortably under $300 a month.

Organise the stack around five jobs#

The reason to think in jobs is that products change names and merge constantly. A category list stays true, and it lets you evaluate a new tool by asking which job it does rather than whether it looks impressive.

Capture#

Getting every enquiry into one place, regardless of how it arrived. This is the least glamorous category and the one most likely to be missing entirely.

The failure mode is obvious once you look for it. Calls go to a mobile, website forms go to an email nobody checks on weekends, Instagram DMs go to whoever has the app, and the text message someone sent your personal number goes nowhere at all. Nothing downstream works properly when enquiries live in four places.

You are looking for a shared inbox that pulls channels together, or an AI agent that sits across them. Expect $0 to $40 a month, because a lot of this is configuration rather than software. Redirecting your website form to a shared address costs nothing and is often the single biggest improvement available.

Qualify#

Sorting the enquiries worth your time from the ones that are not, before you spend an hour on a quote.

A qualifying layer asks two or three things conversationally and attaches the answers to the enquiry. Someone outside your service area gets a polite answer immediately rather than a callback that wastes both of you. Someone with a job starting next week goes to the top of the pile. The dimensions worth asking about, and the phrasing that does not feel like an interrogation, are covered in qualifying leads on your website.

This is usually bundled into whatever answers your enquiries rather than bought separately, so treat it as a feature to check for rather than a line item.

Respond#

Answering fast, at any hour, in a way that books work. This is where most of the budget goes and where most of the return comes from.

An AI agent handling calls, website chat and DMs typically runs $80 to $200 a month for a small business, though the structure matters more than the headline price. Voice usually carries a per-minute charge, chat often carries a per-conversation charge, and the base plan is rarely the whole bill. Ask any vendor what a month at twice your current volume would cost before you sign. The fee structures are broken down in what an AI chatbot costs in New Zealand.

If you only fund one category, fund this one. It is the only category that directly recovers revenue you are currently losing, and it works hardest at the times nobody is at the desk.

Publish#

Producing and scheduling content so the business stays visible without consuming an evening a week.

AI does the drafting and the scheduling well, and it does the judgement about what is worth saying badly. Budget $30 to $60 a month for a scheduling tool with drafting built in, and accept that the ideas still come from you. The honest breakdown of which parts of the social workflow AI handles well sits in the guide to AI social media management.

Resist buying a separate general-purpose writing subscription for this. Drafting is bundled into almost every scheduling tool now, and a standalone writing assistant is the most commonly duplicated line on a small business software bill.

Measure#

Knowing whether any of the above changed anything. This category is nearly free and almost always neglected.

You need analytics on the website, and a place where enquiries and outcomes get recorded so you can see how many became paid work. Many businesses can do this inside tools they already pay for. Expect $0 to $50 a month, plus half an hour of someone's attention monthly. What to actually record is set out in the five metrics that tell you if your AI is working.

What the stack costs#

JobWhat it doesRealistic monthly cost
CapturePulls calls, forms, chat and DMs into one place$0 to $40
QualifySorts real jobs from time-wasters before you quoteUsually bundled with respond
RespondAnswers enquiries instantly, books work, escalates$80 to $200
PublishDrafts and schedules content in your voice$30 to $60
MeasureTracks enquiries, response times and conversions$0 to $50

At the top of every band that comes to $350, and almost nobody needs the top of every band. A typical small service business lands nearer $180 to $250 once the categories that overlap are consolidated.

Compare that against a single recovered job. If your average job is worth $480 and the stack recovers one a month that would otherwise have gone to whoever answered first, the whole thing pays for itself twice over. Whether that holds for your numbers is a fifteen minute calculation, and the method is in working out the ROI of AI automation.

Where the money actually gets wasted#

Overlapping subscriptions are the biggest leak. Tools get bought at different times by different people, and two years later you are paying three vendors for scheduling. List every software charge on your last bank statement and write the job next to each one. Anything without a job, or sharing a job with something else, goes.

The second leak is plan sizing. Vendors price by volume, and it is easy to buy the tier that sounds appropriately serious rather than the tier that matches your traffic. Start on the smallest plan that covers your actual numbers and upgrade when you hit the ceiling, which is a cheaper mistake than the reverse.

The third is buying before defining. A tool with no owner and no job description gets used enthusiastically for three weeks and then forgotten while the subscription runs. Before adding anything, write down the job in one sentence and name the person responsible for it. That habit catches most of the AI mistakes small businesses make before they cost anything.

Build it in the right order#

Do not buy the whole stack at once. Each category takes configuration attention, and doing five at a time means all five get done badly.

Start with capture, because it costs almost nothing and everything else depends on it. Spend a week getting every enquiry channel landing in one place, then look at the volume you were not seeing before. That number often changes what you buy next.

Then add respond, run it in shadow mode for a fortnight where a human approves every reply, and only then let it talk to customers directly. Publish and measure can wait until the first two are stable, which is usually a month or two in.

Almost everything here should be bought rather than built. Custom development earns its cost in a narrow set of circumstances, and the test for whether yours qualifies is in the build versus buy comparison. For how these tools fit alongside the bigger decisions about where to start, the overview of AI solutions for small business covers the sequencing.

Your next step costs nothing. Pull up your last bank statement, list every software subscription, and write the job beside each. The ones with no job answer beside them are your budget for the categories you are missing.

Common questions

How much should a small business spend on AI tools each month?

A service business with two to fifty staff can cover the useful ground for under $300 a month in New Zealand dollars. Most of that goes to whatever answers enquiries, since voice and chat carry usage charges. Everything else tends to sit in the $20 to $60 range per tool.

What AI tools does a small business actually need?

Five jobs rather than five brands. Something to capture enquiries from every channel, something to qualify them, something to respond fast, something to help produce content, and something to measure the result. One paid tool per job is usually enough, and several tools cover two jobs at once.

Where do small businesses waste money on AI tools?

On overlapping subscriptions bought at different times, on plans sized for volume they do not have, and on tools nobody was made responsible for. The second most common waste is paying for a general writing assistant when the same capability is already bundled into a tool they already have.

Let's build your unfair advantage.

Thirty minutes, no pitch deck. Tell me what is leaking and I will tell you whether AI fixes it.