Support automation written for a business of one

Look at the assumptions rather than the recommendations. A queue that can be measured, so that a reduction in it is visible. Multiple people, so that work can be routed and so that somebody notices when an answer is wrong. A manager, so that a report has a reader. A content function, so that writing pages is somebody's job rather than an evening.

Every one of those is a coordination structure, and coordination structures exist because more than one person is involved. Sole traders and one person businesses do not have coordination costs and therefore do not get coordination benefits. Advice about routing, tiering, ownership models, launch communication and stakeholder alignment is not wrong for them. It is about a problem they do not have.

What is left when you take it away is a smaller and more specific set of questions, and it is worth asking them directly rather than translating team advice down. The three that matter are what is actually being saved, where the material is going to come from, and what happens when nobody is watching.

For a team the business case is labour. Fewer questions reaching people means fewer person hours spent, and person hours have a price. That arithmetic does not transfer, because a sole trader's constraint is rarely the total hours available.

The constraint is interruption. The fifteenth message this week asking whether there is parking does not cost fifteen minutes. It costs the interruption, which arrives while you are with a customer, or on a ladder, or mid way through the one piece of concentrated work you had time for today, and takes the rest of the hour with it.

That changes what is worth automating. For a team, the target is the highest volume question, because volume is what costs money. For one person, the target is the question that arrives at the worst moment, which is often not the same question. Messages that come in during your working hours, on your phone, from people who expect a fast reply, are worth more to intercept than a larger number that arrive in the evening when you are going to deal with them in one sitting anyway.

There is no help centre to audit. There is a website with a few pages, most of which are about selling rather than explaining, and everything else lives in your head and in your sent folder. So the work is not editing existing documents. It is writing things down for the first time, which is slower, less pleasant, and the entire project.

The cheapest way through it is not to sit down and write a help centre. It is to answer your next twenty enquiries exactly as you would anyway, and paste each reply into one document as you send it. No editing. Two or three weeks of that and you have your source material, in your own words, covering exactly what people actually ask, and it cost you nothing beyond the pasting.

This is worth doing whether or not you automate anything, which is the real argument for it. That document is what a new employee needs, what a stand in needs when you are ill, and what you will want the day you decide to sell the business or take somebody on. Very few one person businesses have it and nearly all of them would benefit from it.

Escalation in a team means moving a conversation to a different person. For a business of one, it means an interruption to you, on the phone in your pocket, at whatever hour it arrives. There is no tier two. The design question is therefore not who it goes to, it is when you find out.

Askably's whole handover is that it collects a name, an email and a message and sends it to an address the owner nominates, with nothing behind it. For a business of one that is an accurate description of reality rather than a limitation, and the useful decision is which address. Send it to your main inbox and it becomes another stream you are subjected to all day. Send it somewhere separate and it becomes a place you choose to look, twice a day, at times you picked.

The second decision is what deserves to interrupt you immediately, and this is a real design question rather than a setting. In some trades there is a category that genuinely cannot wait until you next look, and it should have a phone number rather than a form. In most, there is not, and the honest thing is a stated reply time you can keep on a bad day rather than a route that pretends to be faster than you are.

A team absorbs bad answers. Somebody sees a strange conversation, somebody gets a confused customer and traces it back, somebody notices the same wrong thing twice. That is error detection, and it is a property of having more than one person.

Alone, there is none. A wrong answer given on Tuesday afternoon while you were on a job is discovered when a customer arrives with the wrong expectation, or is not discovered at all. Nobody is reading conversations in the background. Nobody is going to catch it.

Which inverts one piece of the usual advice. Being narrow and conservative matters more for one person than for a team, not less. A stricter threshold, a shorter list of things it will attempt, a fast handover instead of a strained answer. For a team, a slightly generous setting is recoverable because somebody will notice. For you it is not, so buy the refusals. The cost of a refusal is that you get a message. That is a cost you can see and manage, which is more than can be said for the alternative.

The standard advice is to read conversations regularly, and for a team with a support lead that is a reasonable ask. For one person it competes with invoicing, and it will lose.

So make the unit small enough to actually happen. Once a week, read only the questions it could not answer. On a small site that is a short list, frequently under ten. Read them, pick one, write the answer, put it on the site. Fifteen minutes.

Over a quarter that is a dozen or so answers, written in response to real questions, in priority order determined by real demand. That is a better help centre than most one person businesses ever build deliberately, and it was built in quarter hours. The discipline that works alone is always the smallest repeatable unit, never the thorough periodic audit, because the audit gets scheduled and moved and eventually abandoned and the fifteen minutes does not.

The strongest reason for a one person business to do any of this is not deflection and it is not efficiency. It is the fortnight you are away, or the week you are ill, or the afternoon your phone is off because you are in hospital with somebody.

The alternative in those weeks is an autoresponder that says nothing useful, and a set of enquiries that either wait or go elsewhere. What a business of one needs is for the published answers to keep working while nobody is answering, so that the ordinary questions resolve themselves and only the ones that genuinely need you accumulate.

That reframing makes the content work much easier to justify to yourself, because it stops being a marketing project and becomes continuity planning. It is also more honest. A business of one is a single point of failure, and writing down what you know is the cheapest available mitigation, whether or not anything ever reads it back to a customer.

A short list of things that do not apply, so that you can stop feeling you should be doing them. Containment and deflection percentages, which need a queue you do not have and a baseline you never measured. Tiered escalation. Formal response time targets. A launch plan. Stakeholder sign off. A monthly report, which has no reader.

Two things do apply, and they are the whole of it. Write down the answers, starting with the ones you type most often. And be honest about when you will reply, which is harder alone than in a team, because there is nobody to cover the day you promised something optimistic and then had a job overrun.

That second one is where sole traders most often damage themselves, and it is not a technology problem. A stated reply time of two working days, kept every time, builds more trust than same day kept most of the time. You are the only person who can keep the promise, so make it a promise you can keep on your worst week rather than your best one.