The time a small-business owner loses to admin

Ask a small-business owner what admin costs them. They will answer in sighs, rarely in days. That time is written down nowhere: not on an invoice, not in the accounts, not in a diary. It gets done in the evening, on Saturday, between two jobs, or it does not get done, and then it is a delay.

We measured it at several of our clients, before and after. This article gives no national average: it gives cases, with their figures, and a calculation you can redo for yourself.

A cost nobody writes down anywhere

In a small business, the most expensive person in the company is the owner. They are also, almost always, the one typing invoices, chasing late payers, reconciling transfers and drafting whatever has to be drafted. The work that creates the least value is done by the person whose hour is worth the most.

That is not a failure of organisation, it is a consequence of size. With two or three people, there is nobody else. So the question is not whom to delegate to, but what a system can take over without a person having to come back to it.

Where the days go, in the cases we measured

Invoicing, following contracts, reconciling the bank

A property holding company in Valencia, around ten commercial leases. Every month, the owner issued the invoices, checked the contracts coming up for expiry and reconciled the transfers: two days a month, which happened whenever he had time, which often meant late. Since the AI agent that runs Odoo has issued the invoicing on the 1st and watched the deadlines, he only handles tenants moving in and out. Two to three days a month reclaimed and, something he did not expect, cash: nobody pays late for an invoice that arrived late any more. The full case is published here, and our first article explains the mechanics.

Drafting what has to be drafted

A building firm, a tradesman and his team, inspections at private homes. The inspection is the first step of every job, and its report, long and repetitive, ate into the useful time, the time spent on site. An AI agent now structures the findings and drafts the reports: about 90 % of that load is delegated to it. The figure that matters is elsewhere: the tradesman does more inspections, because he is back where he is useful. The case is here.

Knowing where you stand

A metalwork shop, two people. The owner discovered his annual result after the fact, at closing. He now knows it every week, with the gap to his target. No day is reclaimed here, and yet it may be the heaviest gain: what you discover after the fact can no longer be corrected. The case is here.

Answering at any hour

A premium apart-hotel, several units. Everything rested on the operator: answering guests at any hour, juggling platforms, coordinating cleaning. An AI concierge now welcomes guests day and night, in two languages, and runs the back office. The pilot aims at 80 % less operator time; we write it as a target, not a result, because the roll-out is under way. The case is here.

Two days a month is close to five weeks a year

The trap with admin time is that it comes in small doses. Two days a month is bearable. But a year has twelve of them.

Twenty-four working days is close to five weeks of work. Over a year, the owner of the property holding company spent the equivalent of his holidays invoicing, checking and reconciling. Thirty-six days is more than seven weeks.

What we automate, and what we leave alone

The rule we apply at every client fits in one sentence, written for the property holding company and reused since: the system prepares, checks and alerts; decisions and signatures stay with people. The split, in the cases above:

That split is not caution, it is what makes the system reliable. Software that decided in the owner's place would be one more risk. A system that prepares everything and only waits for a yes is an employee who never sleeps and never forgets.

Where to start

Three questions are enough to spot, in your own business, what should go first.

• What comes back every month, identical? Invoices, reminders, reconciliations. What repeats without judgement is what a system takes over best.
• What gets done « when you have time »? That is what is late. And an admin delay costs twice: the time, then the consequences.
• What do you discover after the fact? A result, an unpaid invoice, a missed deadline. What surprises you is what should have alerted you earlier.

At our clients, the answer to those three questions set the order of work: first Odoo so the data exists in one place, then the AI agent that brings it to life. Never the other way round.

Figures: WEDOOALL client data, anonymised and published with their agreement on our case pages. Measured cases are marked as such; the apart-hotel's 80 % is a pilot target, not a result. The conversion to days per year is arithmetic.