Most businesses don't lose time to big, obvious problems. They lose it to small ones: the report rebuilt every Monday, the details copied from one system into another, the approval that needs chasing twice before anyone signs it off.
Copy. Paste. Repeat. Each task only takes a few minutes, so nobody questions it. But a few minutes several times a week add up quicker than you'd think.
The three-times-a-week rule: if a task happens three or more times a week, follows the same steps each time and uses information that's already digital, it's worth asking whether it could be automated.
Take a single 10-minute task done three times a week. That's 30 minutes a week, or around 23 hours a year across a typical 46-week working year. Multiply that across a team, and across several tasks, and the hours soon add up to weeks.
Time isn't the only cost. Repetitive manual work is where errors creep in, where things get missed when someone's on holiday, and where good people lose interest in their jobs.
Every business is different, but the same kinds of tasks turn up again and again.
Five quick questions separate the good candidates from the rest:
If the answer to the first three is yes, the task is a candidate. The last two tell you how carefully it needs to be designed.
For more on this, please check out our recent blog post, 5 signs your business is ready for AI and automation.
Many of these tasks don't need AI at all. Moving data between systems, sending reminders and routing approvals all follow fixed rules, so a simple workflow handles them reliably and cheaply.
AI earns its place in tasks that involve judgement or language, such as summarising an email, sorting enquiries by type, or drafting a first reply. Even then, a person should approve anything that matters before it goes out. Our Client Strategy Director, Michael Hamer, explored this in his post on why AI-built automation still needs experience.
You don't need a big project to get started. A simple one-week exercise will show you where the time goes:
Start with one task, get it right, then move on to the next.
On Thursday 26 November, we're launching net9.ai with an evening of practical AI and automation at Mad Swans, Farrington Gurney. We'll help the room work out which tasks are worth automating, build a working business app live, and hear from our partners Check Point and Rewst on how to do it safely.
It's free to attend, with a three-course dinner and drinks, and every guest can book a free AI and automation briefing for their business.
What tasks should a small business automate first? Start with tasks that happen three or more times a week, follow the same steps each time and use information that's already digital. Common examples include copying data between systems, chasing invoices, sending follow-up emails and setting up accounts for new starters.
How do I know if a task is worth automating? Check how often it happens, whether the steps are consistent, whether the information is already in a system, what happens if it goes wrong and whether you can measure the benefit. If the first three are true, it's a good candidate.
Do I need AI to automate tasks? Not usually. Rules-based tasks, such as moving data or sending reminders, are best handled by simple workflows. AI is most useful for tasks that involve judgement or language, such as summarising or sorting enquiries.
Is automation secure? It can be, if it's designed properly. Good automation uses the minimum access it needs, is monitored, and keeps a person in the loop for decisions that matter.