Ten working weeks a year, spent on admin

A survey of 167 UK small business owners in trades, property and accounting found an average of eight hours a week going on administration that could largely be automated — quoting, invoicing, chasing payment, rekeying the same details into three systems. That is roughly ten working weeks a year. Seventy-seven per cent of them did it outside working hours, and around 83% had never worked out what it was costing them. (Small sample, commercial research house — directional rather than definitive.)

The figure is worth treating carefully, but its shape will be familiar to anyone running a business of one to nine people. The admin does not stop when the working day does. It follows you home.

The hours are not overhead

In a large company, back-office time is a cost line. In a business of one it is something more expensive: time not spent doing the thing the business is paid for. An electrician doing paperwork is not an electrician. An accountant rekeying data is not advising anyone. Every hour of admin in a micro business is an hour of chargeable expertise that never happened.

That is why the time savings matter more at this size than the efficiency percentages suggest. One survey of 1,000 SME decision-makers put the average saving among those using AI at 5.2 hours a week. (Opinium via Enterprise Nation — vendor-commissioned; treat as indicative.) The two surveys measure different populations and shouldn’t be set directly against each other, but they describe the same territory: a working week with several hours of it spoken for, and a plausible route to getting some of it back.

Trade body evidence points the same way without putting a number on the clock. The Federation of Small Businesses reports 59% of small firms using AI see productivity gains and 24% report higher revenue. The revenue figure is the telling one: it suggests the recovered time is going into work that bills.

What it adds up to

Government has done the arithmetic at the national scale. The DBT SME Digital Adoption Taskforce estimated that a 1% productivity uplift across UK SMEs would add around £94bn a year to GDP, citing Enterprise Research Centre findings of 7–18% firm-level productivity gains per technology adopted.

For an owner, though, the unit is not GDP. It is whether the quotes go out on Thursday afternoon instead of Sunday night.

The reason it doesn’t happen

There is a trap in this. Among the barriers to AI adoption recorded by DBT/Ipsos, 21% cite a lack of time or capacity to explore the options. (DBT/Ipsos, 2025) The firms losing the most hours to admin are the least able to spare the hours it takes to fix it. Under-investment follows — only 13% of UK SMEs plan to increase technology investment, against an 18% European average.

What the evidence does not show is any lack of appetite once the first system is working. ONS analysis indicates that 17% of adopters with 0–9 employees use AI extensively, against 9% of adopters in firms with 250 or more staff. The smallest businesses that commit, commit hardest. Getting the first hour back is the difficult part.