11For your size
AI implementation for a small business, with nothing to pay up front
You are the owner. You are also the person who answers the phone, chases the invoice, and fixes the printer. That is not a small version of a big company. It is a different problem, and it needs a different answer.
01
When one person holds three jobs
In a company of two hundred people, three different people decide whether to do this. A budget holder signs it. A department runs it. Somebody in IT keeps it alive.
In your business, that is you. You are the buyer, the user, and the IT department.
Two things follow from that, and they pull in opposite directions.
The good one: you can decide in a single conversation. There is no committee, no business case, no waiting for the next quarter. If it makes sense to you on a Tuesday, it can start on a Wednesday.
The hard one: if you do not use it, nobody else will. In a larger firm a new way of working survives because a department adopts it. Here it survives because one person changed a habit. So everything we build for a business this size has to fit into what you already do all day. If it needs you to open a new tab and remember a new routine, it will be dead in six weeks and we will have wasted your time and ours.
That is why the first question we ask an owner is not what you want to automate. It is what you actually do between eight and six.
02
Off-the-shelf tools only, no custom platform
Here is the rule we hold to at this size: we do not build you a bespoke system.
Custom software needs someone to look after it. Something changes, an interface moves, a provider updates a rule, and a person has to fix it that afternoon. In a company with an IT team, that person exists. In your business, that person is you, and you have a business to run.
So we build on ordinary tools you can buy off the shelf, with a support line you can call and a help page you can read. In practice that usually means the office suite you already have, the accounting package you already file with, the booking or till system your customers already use, the shared inbox, and a connector that joins them up. Almost all of it you are already paying for, and most of it can already do more than it is being asked to do.
If we leave and you never speak to us again, everything we set up keeps running, and any competent person can pick it up. That is a deliberate constraint. It costs a little capability. It buys you the ability to walk away.
03
The two places it pays back first
At this size the answer is nearly always the same two, in this order.
- 01
First, intake. The enquiry that arrives when nobody is there. A missed call at a big company goes into a queue. A missed call at your business goes nowhere. There is no second line, no overflow team, and no evening shift. So the work here is simple: every call, form, message and walk-up gets an answer straight away, at any hour, with your real availability and your real prices, and a booking made if that is what they want. You wake up to a booked diary instead of three voicemails and a maybe.
- 02
Second, admin. The hour you spend after closing, doing the part of the job nobody pays you for. Quotes retyped into invoices. Receipts photographed and coded. Appointment reminders sent one at a time. The same four questions answered by hand. Stock or supplies reordered from memory. It is unglamorous, it is boring, and it is where the evening goes.
We start there rather than with anything cleverer because those two are countable, they show a result inside weeks, and they hand you back the only thing you cannot buy more of.
04
What the gain looks like without cutting staff
This is the part most AI sales copy gets wrong for a business your size, so we want to be plain about it. At a large company, automation shows up as cost taken out. Fewer hours paid for the same work. That arithmetic does not work here, because you were not going to lay anybody off. In a business of one, or four, or nine, there is nobody spare. So the gain arrives in two other shapes.
- 01
Work you were dropping, now caught. The Saturday enquiry that went cold. The follow up you meant to send on Thursday. The invoice raised nine days late. None of that is a saving. It is money that was already yours and was falling out of the bottom of the week.
- 02
Capacity you did not have to hire for. The next ten customers absorbed without a tenth person. For a business at the edge of what its owner can personally handle, that is the whole ballgame, and it is worth more than any hourly saving on a spreadsheet.
Both are measurable, and both are measured against what the business was already doing before anything changed. That matters, because that measurement is the whole basis of how Praxis gets paid.
05
The first ninety days
These are the shapes this usually takes, not promises. They move with how the business is put together and how quickly you can get us what we need.
- 01
Weeks 1 to 2.
A free call, then a look at the real work. For a business this size that is usually a couple of days rather than a fortnight, because there are fewer places for the work to hide. We want a normal week, not a good one.
- 02
Weeks 3 to 6.
The first piece goes live. One thing, not a programme. Almost always intake, because it is the fastest to stand up and the easiest to see working. It runs beside your old way, not instead of it, until you trust it.
- 03
Weeks 6 to 10.
The admin piece. Now the two are joined, so the enquiry that came in on Saturday becomes a booking, becomes a job, becomes an invoice, without you retyping anything between those steps.
- 04
Weeks 10 to 13.
The numbers the business was already producing get read again, the same way they were read at the start. You get written instructions in plain language: what it does, what to check each week, what to do if it stops, and how to switch it off.
06
If the first piece does not work, it is week five and one thing is switched off. Nothing here is built so large that stopping is expensive.
07
What a small business should not try to automate
The honest half of this page. Some of it is a bad idea at your size even though it works fine at a bigger one.
- 01
Anything that happens a handful of times a month. Automation earns its keep on repetition. If a job happens four times a month and takes ten minutes, leave it alone. Doing it by hand is cheaper than owning a machine that does it.
- 02
The reason people buy from you rather than the big competitor. If customers come to you because you personally know them and answer personally, do not put a machine in the middle of that. Automate what happens before that conversation and after it, never the conversation itself.
- 03
Pricing a job you have to look at. Anything that depends on your eye, your judgement, or walking into a room and seeing the state of it. A machine can prepare the quote. It should not decide it.
- 04
Anything that pretends to be you. Automated replies say plainly that they are automated and give a direct route to a person. At your size your name is the business, and being caught faking your own voice costs more than every hour it saved.
- 05
A process that is broken rather than slow. If the way the job runs is wrong, automating it just makes the wrong thing happen faster. That is a free fix and it should happen first.
- 06
Anything that leaves you unable to answer a customer. If the machine holds information you cannot see or change yourself, you will one day be on the phone to a customer, unable to explain what happened. Do not accept that trade.
08
What it costs, and what it does not
There is no bill for the call, no bill for the look at the work, and no bill for the build. The fee is a share of the measured gain. If there is no measured gain, there is no share.
That is the fee, and it is worth naming separately what you do spend regardless of how it turns out. You spend your time: a few hours in the first fortnight and less after that. You spend access to how the business really runs, which for a lot of owners is the uncomfortable part. You spend your attention on changing a habit you have had for years. And you spend a little patience while the new way runs beside the old one.
That is real, it is not nothing, and we would rather you knew it before the call than after it. How the starting numbers are agreed and how the share is measured is set out on how you pay.
09
Where to look next
Intake and admin are the two functions this page is really about, and each has a page of its own: answering customers day and night and the admin behind the business. If you want the whole offer rather than the small business version of it, that is the place to go.
Questions
What should you know before the first call?
- Can a small business with no IT staff actually run AI automation?
- Yes, provided it is built on off-the-shelf tools rather than custom software. Bespoke systems need somebody to maintain them, and in a small business that person would be the owner. Ordinary business tools with a support line and documentation can be joined together, handed over with written instructions, and kept running by whoever is there.
- What should a small business automate first?
- Intake and admin, in that order. Intake means every call, message and enquiry getting a real answer straight away at any hour, because a missed enquiry at a small business goes nowhere rather than into a queue. Admin means the after-hours hour: quotes turned into invoices, receipts coded, reminders sent, routine questions answered.
- Does AI automation only pay off if staff are cut?
- No, and at a small business it usually is not about staff at all. The gain shows up as work that was being dropped and is now caught, such as an enquiry that arrived on a weekend, and as capacity to take on more customers without hiring. Both are measured against what the business was already producing before anything changed.
- What should a small business avoid automating?
- Anything that happens only a few times a month, anything that depends on the owner's own judgement, and the personal relationship that makes customers choose a small business over a large one. Automate what happens before and after that conversation, not the conversation. Also avoid automating a process that is broken rather than slow, since that only makes the wrong result arrive faster.
- How long before a small business sees anything working?
- The first piece is usually live within about six weeks, because it is one thing rather than a programme, and it runs alongside the existing way of working until it is trusted. That is the shape this normally takes rather than a fixed commitment, and it moves with how the business is set up.
- Is there any cost to find out whether this is worth doing?
- No. The call costs nothing and the look at how the business runs costs nothing. The fee is a share of the measured gain once something has been built and measured against numbers agreed in advance. If there is no measured gain, there is no share.
Where this leads
This page is one part of the whole offer: the whole offer.
Start with a free consultation
Tell us what you do between eight and six and which hour of the day disappears into admin. We will tell you which of it can be taken over, what should be left alone at your size, and what we would set up first. The call costs nothing and ends with a plain list either way.
No obligation · a scoping conversation first